{ "generated_at": "2026-07-10", "from_date": "2025-01-01", "to_date": "2026-07-10", "query_totals": { "agent-memory": 369, "long-term-memory": 109, "episodic-memory": 60, "procedural-memory": 10, "memory-evaluation": 42 }, "unique_papers": 516, "papers": [ { "arxiv_id": "2501.13956", "title": "Zep: A Temporal Knowledge Graph Architecture for Agent Memory", "abstract": "We introduce Zep, a novel memory layer service for AI agents that outperforms the current state-of-the-art system, MemGPT, in the Deep Memory Retrieval (DMR) benchmark. Additionally, Zep excels in more comprehensive and challenging evaluations than DMR that better reflect real-world enterprise use cases. While existing retrieval-augmented generation (RAG) frameworks for large language model (LLM)-based agents are limited to static document retrieval, enterprise applications demand dynamic knowledge integration from diverse sources including ongoing conversations and business data. Zep addresses this fundamental limitation through its core component Graphiti -- a temporally-aware knowledge graph engine that dynamically synthesizes both unstructured conversational data and structured business data while maintaining historical relationships. In the DMR benchmark, which the MemGPT team established as their primary evaluation metric, Zep demonstrates superior performance (94.8% vs 93.4%). Beyond DMR, Zep's capabilities are further validated through the more challenging LongMemEval benchmark, which better reflects enterprise use cases through complex temporal reasoning tasks. In this evaluation, Zep achieves substantial results with accuracy improvements of up to 18.5% while simultaneously reducing response latency by 90% compared to baseline implementations. These results are particularly pronounced in enterprise-critical tasks such as cross-session information synthesis and long-term context maintenance, demonstrating Zep's effectiveness for deployment in real-world applications.", "published": "2025-01-20", "updated": "2025-01-20", "authors": [ "Preston Rasmussen", "Pavlo Paliychuk", "Travis Beauvais", "Jack Ryan", "Daniel Chalef" ], "categories": [ "cs.CL", "cs.AI", "cs.IR" ], "url": "https://arxiv.org/abs/2501.13956", "pdf_url": "https://arxiv.org/pdf/2501.13956", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2501.12485", "title": "R2D2: Remembering, Replaying and Dynamic Decision Making with a Reflective Agentic Memory", "abstract": "The proliferation of web agents necessitates advanced navigation and interaction strategies within complex web environments. Current models often struggle with efficient navigation and action execution due to limited visibility and understanding of web structures. Our proposed R2D2 framework addresses these challenges by integrating two paradigms: Remember and Reflect. The Remember paradigm uses a replay buffer that aids agents in reconstructing the web environment dynamically, thus enabling the formulation of a detailed \"map\" of previously visited pages. This helps in reducing navigational errors and optimizing the decision-making process during web interactions. Conversely, the Reflect paradigm allows agents to learn from past mistakes by providing a mechanism for error analysis and strategy refinement, enhancing overall task performance. We evaluate R2D2 using the WebArena benchmark, demonstrating substantial improvements over existing methods, including a 50% reduction in navigation errors and a threefold increase in task completion rates. Our findings suggest that a combination of memory-enhanced navigation and reflective learning promisingly advances the capabilities of web agents, potentially benefiting various applications such as automated customer service and personal digital assistants.", "published": "2025-01-21", "updated": "2025-07-22", "authors": [ "Tenghao Huang", "Kinjal Basu", "Ibrahim Abdelaziz", "Pavan Kapanipathi", "Jonathan May", "Muhao Chen" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2501.12485", "pdf_url": "https://arxiv.org/pdf/2501.12485", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2501.18160", "title": "RepoAudit: An Autonomous LLM-Agent for Repository-Level Code Auditing", "abstract": "Code auditing is the process of reviewing code with the aim of identifying bugs. Large Language Models (LLMs) have demonstrated promising capabilities for this task without requiring compilation, while also supporting user-friendly customization. However, auditing a code repository with LLMs poses significant challenges: limited context windows and hallucinations can degrade the quality of bug reports, and analyzing large-scale repositories incurs substantial time and token costs, hindering efficiency and scalability. This work introduces an LLM-based agent, RepoAudit, designed to perform autonomous repository-level code auditing. Equipped with agent memory, RepoAudit explores the codebase on demand by analyzing data-flow facts along feasible program paths within individual functions. It further incorporates a validator module to mitigate hallucinations by verifying data-flow facts and checking the satisfiability of path conditions associated with potential bugs, thereby reducing false positives. RepoAudit detects 40 true bugs across 15 real-world benchmark projects with a precision of 78.43%, requiring on average only 0.44 hours and $2.54 per project. Also, it detects 185 new bugs in high-profile projects, among which 174 have been confirmed or fixed. We have open-sourced RepoAudit at https://github.com/PurCL/RepoAudit.", "published": "2025-01-30", "updated": "2025-05-29", "authors": [ "Jinyao Guo", "Chengpeng Wang", "Xiangzhe Xu", "Zian Su", "Xiangyu Zhang" ], "categories": [ "cs.SE", "cs.PL" ], "url": "https://arxiv.org/abs/2501.18160", "pdf_url": "https://arxiv.org/pdf/2501.18160", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2502.06975", "title": "Position: Episodic Memory is the Missing Piece for Long-Term LLM Agents", "abstract": "As Large Language Models (LLMs) evolve from text-completion tools into fully fledged agents operating in dynamic environments, they must address the challenge of continually learning and retaining long-term knowledge. Many biological systems solve these challenges with episodic memory, which supports single-shot learning of instance-specific contexts. Inspired by this, we present an episodic memory framework for LLM agents, centered around five key properties of episodic memory that underlie adaptive and context-sensitive behavior. With various research efforts already partially covering these properties, this position paper argues that now is the right time for an explicit, integrated focus on episodic memory to catalyze the development of long-term agents. To this end, we outline a roadmap that unites several research directions under the goal to support all five properties of episodic memory for more efficient long-term LLM agents.", "published": "2025-02-10", "updated": "2025-02-10", "authors": [ "Mathis Pink", "Qinyuan Wu", "Vy Ai Vo", "Javier Turek", "Jianing Mu", "Alexander Huth", "Mariya Toneva" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2502.06975", "pdf_url": "https://arxiv.org/pdf/2502.06975", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2502.10177", "title": "STMA: A Spatio-Temporal Memory Agent for Long-Horizon Embodied Task Planning", "abstract": "A key objective of embodied intelligence is enabling agents to perform long-horizon tasks in dynamic environments while maintaining robust decision-making and adaptability. To achieve this goal, we propose the Spatio-Temporal Memory Agent (STMA), a novel framework designed to enhance task planning and execution by integrating spatio-temporal memory. STMA is built upon three critical components: (1) a spatio-temporal memory module that captures historical and environmental changes in real time, (2) a dynamic knowledge graph that facilitates adaptive spatial reasoning, and (3) a planner-critic mechanism that iteratively refines task strategies. We evaluate STMA in the TextWorld environment on 32 tasks, involving multi-step planning and exploration under varying levels of complexity. Experimental results demonstrate that STMA achieves a 31.25% improvement in success rate and a 24.7% increase in average score compared to the state-of-the-art model. The results highlight the effectiveness of spatio-temporal memory in advancing the memory capabilities of embodied agents.", "published": "2025-02-14", "updated": "2025-03-02", "authors": [ "Mingcong Lei", "Yiming Zhao", "Ge Wang", "Zhixin Mai", "Shuguang Cui", "Yatong Han", "Jinke Ren" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2502.10177", "pdf_url": "https://arxiv.org/pdf/2502.10177", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2502.10550", "title": "Memory, Benchmark & Robots: A Benchmark for Solving Complex Tasks with Reinforcement Learning", "abstract": "Memory is crucial for enabling agents to tackle complex tasks with temporal and spatial dependencies. While many reinforcement learning (RL) algorithms incorporate memory, the field lacks a universal benchmark to assess an agent's memory capabilities across diverse scenarios. This gap is particularly evident in tabletop robotic manipulation, where memory is essential for solving tasks with partial observability and ensuring robust performance, yet no standardized benchmarks exist. To address this, we introduce MIKASA (Memory-Intensive Skills Assessment Suite for Agents), a comprehensive benchmark for memory RL, with three key contributions: (1) we propose a comprehensive classification framework for memory-intensive RL tasks, (2) we collect MIKASA-Base -- a unified benchmark that enables systematic evaluation of memory-enhanced agents across diverse scenarios, and (3) we develop MIKASA-Robo (pip install mikasa-robo-suite) -- a novel benchmark of 32 carefully designed memory-intensive tasks that assess memory capabilities in tabletop robotic manipulation. Our work introduces a unified framework to advance memory RL research, enabling more robust systems for real-world use. MIKASA is available at https://tinyurl.com/membenchrobots.", "published": "2025-02-14", "updated": "2026-03-04", "authors": [ "Egor Cherepanov", "Nikita Kachaev", "Alexey K. Kovalev", "Aleksandr I. Panov" ], "categories": [ "cs.LG", "cs.AI", "cs.RO" ], "url": "https://arxiv.org/abs/2502.10550", "pdf_url": "https://arxiv.org/pdf/2502.10550", "matched_queries": [ "agent-memory", "memory-evaluation" ] }, { "arxiv_id": "2502.12110", "title": "A-MEM: Agentic Memory for LLM Agents", "abstract": "While large language model (LLM) agents can effectively use external tools for complex real-world tasks, they require memory systems to leverage historical experiences. Current memory systems enable basic storage and retrieval but lack sophisticated memory organization, despite recent attempts to incorporate graph databases. Moreover, these systems' fixed operations and structures limit their adaptability across diverse tasks. To address this limitation, this paper proposes a novel agentic memory system for LLM agents that can dynamically organize memories in an agentic way. Following the basic principles of the Zettelkasten method, we designed our memory system to create interconnected knowledge networks through dynamic indexing and linking. When a new memory is added, we generate a comprehensive note containing multiple structured attributes, including contextual descriptions, keywords, and tags. The system then analyzes historical memories to identify relevant connections, establishing links where meaningful similarities exist. Additionally, this process enables memory evolution - as new memories are integrated, they can trigger updates to the contextual representations and attributes of existing historical memories, allowing the memory network to continuously refine its understanding. Our approach combines the structured organization principles of Zettelkasten with the flexibility of agent-driven decision making, allowing for more adaptive and context-aware memory management. Empirical experiments on six foundation models show superior improvement against existing SOTA baselines. The source code for evaluating performance is available at https://github.com/WujiangXu/A-mem, while the source code of the agentic memory system is available at https://github.com/WujiangXu/A-mem-sys.", "published": "2025-02-17", "updated": "2025-10-08", "authors": [ "Wujiang Xu", "Zujie Liang", "Kai Mei", "Hang Gao", "Juntao Tan", "Yongfeng Zhang" ], "categories": [ "cs.CL", "cs.HC" ], "url": "https://arxiv.org/abs/2502.12110", "pdf_url": "https://arxiv.org/pdf/2502.12110", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2502.13172", "title": "Unveiling Privacy Risks in LLM Agent Memory", "abstract": "Large Language Model (LLM) agents have become increasingly prevalent across various real-world applications. They enhance decision-making by storing private user-agent interactions in the memory module for demonstrations, introducing new privacy risks for LLM agents. In this work, we systematically investigate the vulnerability of LLM agents to our proposed Memory EXTRaction Attack (MEXTRA) under a black-box setting. To extract private information from memory, we propose an effective attacking prompt design and an automated prompt generation method based on different levels of knowledge about the LLM agent. Experiments on two representative agents demonstrate the effectiveness of MEXTRA. Moreover, we explore key factors influencing memory leakage from both the agent designer's and the attacker's perspectives. Our findings highlight the urgent need for effective memory safeguards in LLM agent design and deployment.", "published": "2025-02-17", "updated": "2025-06-03", "authors": [ "Bo Wang", "Weiyi He", "Shenglai Zeng", "Zhen Xiang", "Yue Xing", "Jiliang Tang", "Pengfei He" ], "categories": [ "cs.CR", "cs.AI" ], "url": "https://arxiv.org/abs/2502.13172", "pdf_url": "https://arxiv.org/pdf/2502.13172", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2502.16090", "title": "Echo: A Large Language Model with Temporal Episodic Memory", "abstract": "Research on large language models (LLMs) has shown remarkable performance in domains such as mathematics, programming, and literary creation. However, most studies have focused on semantic memory-based question answering, neglecting LLMs' potential to handle episodic memory (EM)-related queries. This oversight has led to suboptimal performance in applications requiring EM, including emotional companionship, personal AI assistants, and AI teachers. To address this gap, we introduce Echo, a LLM enhanced with temporal episodic memory. We propose a Multi-Agent Data Generation Framework that guides the model in generating multi-turn, complex scenario episodic memory dialogue data (EM-Train). Temporal information is innovatively incorporated into the LLM training process, and Echo is trained using the EM-Train. Furthermore, We develop an EM-Test benchmark specifically designed to evaluate LLMs' episodic memory capabilities. The EM-Test assesses performance across various time spans and difficulty levels, providing a comprehensive evaluation of multi-turn episodic memory dialogues. Our experiments demonstrate that Echo significantly outperforms state-of-the-art LLMs on EM-Test. Additionally, a qualitative analysis reveals Echo's potential to exhibit human-like episodic memory capabilities. We will open-source all datasets, code, and model weights.", "published": "2025-02-22", "updated": "2025-02-22", "authors": [ "WenTao Liu", "Ruohua Zhang", "Aimin Zhou", "Feng Gao", "JiaLi Liu" ], "categories": [ "cs.CL", "cs.AI", "cs.LG" ], "url": "https://arxiv.org/abs/2502.16090", "pdf_url": "https://arxiv.org/pdf/2502.16090", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2502.19145", "title": "Multi-Agent Security Tax: Trading Off Security and Collaboration Capabilities in Multi-Agent Systems", "abstract": "As AI agents are increasingly adopted to collaborate on complex objectives, ensuring the security of autonomous multi-agent systems becomes crucial. We develop simulations of agents collaborating on shared objectives to study these security risks and security trade-offs. We focus on scenarios where an attacker compromises one agent, using it to steer the entire system toward misaligned outcomes by corrupting other agents. In this context, we observe infectious malicious prompts - the multi-hop spreading of malicious instructions. To mitigate this risk, we evaluated several strategies: two \"vaccination\" approaches that insert false memories of safely handling malicious input into the agents' memory stream, and two versions of a generic safety instruction strategy. While these defenses reduce the spread and fulfillment of malicious instructions in our experiments, they tend to decrease collaboration capability in the agent network. Our findings illustrate potential trade-off between security and collaborative efficiency in multi-agent systems, providing insights for designing more secure yet effective AI collaborations.", "published": "2025-02-26", "updated": "2025-06-04", "authors": [ "Pierre Peigne-Lefebvre", "Mikolaj Kniejski", "Filip Sondej", "Matthieu David", "Jason Hoelscher-Obermaier", "Christian Schroeder de Witt", "Esben Kran" ], "categories": [ "cs.AI", "cs.MA" ], "url": "https://arxiv.org/abs/2502.19145", "pdf_url": "https://arxiv.org/pdf/2502.19145", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2503.03704", "title": "Memory Injection Attacks on LLM Agents via Query-Only Interaction", "abstract": "Agents powered by large language models (LLMs) have demonstrated strong capabilities in a wide range of complex, real-world applications. However, LLM agents with a compromised memory bank may easily produce harmful outputs when the past records retrieved for demonstration are malicious. In this paper, we propose a novel Memory INJection Attack, MINJA, without assuming that the attacker can directly modify the memory bank of the agent. The attacker injects malicious records into the memory bank by only interacting with the agent via queries and output observations. These malicious records are designed to elicit a sequence of malicious reasoning steps corresponding to a different target query during the agent's execution of the victim user's query. Specifically, we introduce a sequence of bridging steps to link victim queries to the malicious reasoning steps. During the memory injection, we propose an indication prompt that guides the agent to autonomously generate similar bridging steps, with a progressive shortening strategy that gradually removes the indication prompt, such that the malicious record will be easily retrieved when processing later victim queries. Our extensive experiments across diverse agents demonstrate the effectiveness of MINJA in compromising agent memory. With minimal requirements for execution, MINJA enables any user to influence agent memory, highlighting the risk.", "published": "2025-03-05", "updated": "2026-02-12", "authors": [ "Shen Dong", "Shaochen Xu", "Pengfei He", "Yige Li", "Jiliang Tang", "Tianming Liu", "Hui Liu", "Zhen Xiang" ], "categories": [ "cs.LG" ], "url": "https://arxiv.org/abs/2503.03704", "pdf_url": "https://arxiv.org/pdf/2503.03704", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2503.04392", "title": "AgentSafe: Safeguarding Large Language Model-based Multi-agent Systems via Hierarchical Data Management", "abstract": "Large Language Model based multi-agent systems are revolutionizing autonomous communication and collaboration, yet they remain vulnerable to security threats like unauthorized access and data breaches. To address this, we introduce AgentSafe, a novel framework that enhances MAS security through hierarchical information management and memory protection. AgentSafe classifies information by security levels, restricting sensitive data access to authorized agents. AgentSafe incorporates two components: ThreatSieve, which secures communication by verifying information authority and preventing impersonation, and HierarCache, an adaptive memory management system that defends against unauthorized access and malicious poisoning, representing the first systematic defense for agent memory. Experiments across various LLMs show that AgentSafe significantly boosts system resilience, achieving defense success rates above 80% under adversarial conditions. Additionally, AgentSafe demonstrates scalability, maintaining robust performance as agent numbers and information complexity grow. Results underscore effectiveness of AgentSafe in securing MAS and its potential for real-world application.", "published": "2025-03-06", "updated": "2025-07-08", "authors": [ "Junyuan Mao", "Fanci Meng", "Yifan Duan", "Miao Yu", "Xiaojun Jia", "Junfeng Fang", "Yuxuan Liang", "Kun Wang", "Qingsong Wen" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2503.04392", "pdf_url": "https://arxiv.org/pdf/2503.04392", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2503.21760", "title": "MemInsight: Autonomous Memory Augmentation for LLM Agents", "abstract": "Large language model (LLM) agents have evolved to intelligently process information, make decisions, and interact with users or tools. A key capability is the integration of long-term memory capabilities, enabling these agents to draw upon historical interactions and knowledge. However, the growing memory size and need for semantic structuring pose significant challenges. In this work, we propose an autonomous memory augmentation approach, MemInsight, to enhance semantic data representation and retrieval mechanisms. By leveraging autonomous augmentation to historical interactions, LLM agents are shown to deliver more accurate and contextualized responses. We empirically validate the efficacy of our proposed approach in three task scenarios; conversational recommendation, question answering and event summarization. On the LLM-REDIAL dataset, MemInsight boosts persuasiveness of recommendations by up to 14%. Moreover, it outperforms a RAG baseline by 34% in recall for LoCoMo retrieval. Our empirical results show the potential of MemInsight to enhance the contextual performance of LLM agents across multiple tasks.", "published": "2025-03-27", "updated": "2025-07-31", "authors": [ "Rana Salama", "Jason Cai", "Michelle Yuan", "Anna Currey", "Monica Sunkara", "Yi Zhang", "Yassine Benajiba" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2503.21760", "pdf_url": "https://arxiv.org/pdf/2503.21760", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2503.23514", "title": "If an LLM Were a Character, Would It Know Its Own Story? Evaluating Lifelong Learning in LLMs", "abstract": "Large language models (LLMs) can carry out human-like dialogue, but unlike humans, they are stateless due to the superposition property. However, during multi-turn, multi-agent interactions, LLMs begin to exhibit consistent, character-like behaviors, hinting at a form of emergent lifelong learning. Despite this, existing benchmarks often fail to capture these dynamics, primarily focusing on static, open-ended evaluations. To address this gap, we introduce LIFESTATE-BENCH, a benchmark designed to assess lifelong learning in LLMs. It features two episodic datasets: Hamlet and a synthetic script collection, rich in narrative structure and character interactions. Our fact checking evaluation probes models' self-awareness, episodic memory retrieval, and relationship tracking, across both parametric and non-parametric approaches. Experiments on models like Llama3.1-8B, GPT-4-turbo, and DeepSeek R1, we demonstrate that nonparametric methods significantly outperform parametric ones in managing stateful learning. However, all models exhibit challenges with catastrophic forgetting as interactions extend, highlighting the need for further advancements in lifelong learning.", "published": "2025-03-30", "updated": "2026-04-13", "authors": [ "Siqi Fan", "Xiusheng Huang", "Yiqun Yao", "Xuezhi Fang", "Kang Liu", "Peng Han", "Shuo Shang", "Aixin Sun", "Yequan Wang" ], "categories": [ "cs.CL", "cs.AI" ], "url": "https://arxiv.org/abs/2503.23514", "pdf_url": "https://arxiv.org/pdf/2503.23514", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2504.09283", "title": "Semantic Commit: Helping Users Update Intent Specifications for AI Memory at Scale", "abstract": "How do we update AI memory of user intent as intent changes? We consider how an AI interface may assist the integration of new information into a repository of natural language data. Inspired by software engineering concepts like impact analysis, we develop methods and a UI for managing semantic changes with non-local effects, which we call \"semantic conflict resolution.\" The user commits new intent to a project -- makes a \"semantic commit\" -- and the AI helps the user detect and resolve semantic conflicts within a store of existing information representing their intent (an \"intent specification\"). We develop an interface, SemanticCommit, to better understand how users resolve conflicts when updating intent specifications such as Cursor Rules and game design documents. A knowledge graph-based RAG pipeline drives conflict detection, while LLMs assist in suggesting resolutions. We evaluate our technique on an initial benchmark. Then, we report a 12 user within-subjects study of SemanticCommit for two task domains -- game design documents, and AI agent memory in the style of ChatGPT memories -- where users integrated new information into an existing list. Half of our participants adopted a workflow of impact analysis, where they would first flag conflicts without AI revisions then resolve conflicts locally, despite having access to a global revision feature. We argue that AI agent interfaces, such as software IDEs like Cursor and Windsurf, should provide affordances for impact analysis and help users validate AI retrieval independently from generation. Our work speaks to how AI agent designers should think about updating memory as a process that involves human feedback and decision-making.", "published": "2025-04-12", "updated": "2025-04-12", "authors": [ "Priyan Vaithilingam", "Munyeong Kim", "Frida-Cecilia Acosta-Parenteau", "Daniel Lee", "Amine Mhedhbi", "Elena L. Glassman", "Ian Arawjo" ], "categories": [ "cs.HC", "cs.AI" ], "url": "https://arxiv.org/abs/2504.09283", "pdf_url": "https://arxiv.org/pdf/2504.09283", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2504.16946", "title": "MobileCity: An Efficient Framework for Large-Scale Urban Behavior Simulation", "abstract": "Generative agents offer promising capabilities for simulating realistic urban behaviors. However, existing methods oversimplify transportation choices, rely heavily on static agent profiles leading to behavioral homogenization, and inherit prohibitive computational costs. To address these limitations, we present MobileCity, a lightweight simulation platform designed to model realistic urban mobility with high computational efficiency. We introduce a comprehensive transportation system with multiple transport modes, and collect questionnaire data from respondents to construct agent profiles. To enable scalable simulation, agents perform action selection within a pre-generated action space and uses local models for efficient agent memory generation. Through extensive micro and macro-level evaluations on 4,000 agents, we demonstrate that MobileCity generates more realistic urban behaviors than baselines while maintaining computational efficiency. We further explore practical applications such as predicting movement patterns and analyzing demographic trends in transportation preferences. Our code is publicly available at https://github.com/Tony-Yip/MobileCity.", "published": "2025-04-18", "updated": "2026-01-26", "authors": [ "Xiaotong Ye", "Nicolas Bougie", "Toshihiko Yamasaki", "Narimasa Watanabe" ], "categories": [ "cs.SI", "cs.AI" ], "url": "https://arxiv.org/abs/2504.16946", "pdf_url": "https://arxiv.org/pdf/2504.16946", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2504.15263", "title": "Interpretable Locomotion Prediction in Construction Using a Memory-Driven LLM Agent With Chain-of-Thought Reasoning", "abstract": "Construction tasks are inherently unpredictable, with dynamic environments and safety-critical demands posing significant risks to workers. Exoskeletons offer potential assistance but falter without accurate intent recognition across diverse locomotion modes. This paper presents a locomotion prediction agent leveraging Large Language Models (LLMs) augmented with memory systems, aimed at improving exoskeleton assistance in such settings. Using multimodal inputs - spoken commands and visual data from smart glasses - the agent integrates a Perception Module, Short-Term Memory (STM), Long-Term Memory (LTM), and Refinement Module to predict locomotion modes effectively. Evaluation reveals a baseline weighted F1-score of 0.73 without memory, rising to 0.81 with STM, and reaching 0.90 with both STM and LTM, excelling with vague and safety-critical commands. Calibration metrics, including a Brier Score drop from 0.244 to 0.090 and ECE from 0.222 to 0.044, affirm improved reliability. This framework supports safer, high-level human-exoskeleton collaboration, with promise for adaptive assistive systems in dynamic industries.", "published": "2025-04-21", "updated": "2025-04-21", "authors": [ "Ehsan Ahmadi", "Chao Wang" ], "categories": [ "cs.RO" ], "url": "https://arxiv.org/abs/2504.15263", "pdf_url": "https://arxiv.org/pdf/2504.15263", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2504.20117", "title": "ResearchCodeAgent: An LLM Multi-Agent System for Automated Codification of Research Methodologies", "abstract": "In this paper we introduce ResearchCodeAgent, a novel multi-agent system leveraging large language models (LLMs) agents to automate the codification of research methodologies described in machine learning literature. The system bridges the gap between high-level research concepts and their practical implementation, allowing researchers auto-generating code of existing research papers for benchmarking or building on top-of existing methods specified in the literature with availability of partial or complete starter code. ResearchCodeAgent employs a flexible agent architecture with a comprehensive action suite, enabling context-aware interactions with the research environment. The system incorporates a dynamic planning mechanism, utilizing both short and long-term memory to adapt its approach iteratively. We evaluate ResearchCodeAgent on three distinct machine learning tasks with distinct task complexity and representing different parts of the ML pipeline: data augmentation, optimization, and data batching. Our results demonstrate the system's effectiveness and generalizability, with 46.9% of generated code being high-quality and error-free, and 25% showing performance improvements over baseline implementations. Empirical analysis shows an average reduction of 57.9% in coding time compared to manual implementation. We observe higher gains for more complex tasks. ResearchCodeAgent represents a significant step towards automating the research implementation process, potentially accelerating the pace of machine learning research.", "published": "2025-04-28", "updated": "2025-05-03", "authors": [ "Shubham Gandhi", "Dhruv Shah", "Manasi Patwardhan", "Lovekesh Vig", "Gautam Shroff" ], "categories": [ "cs.SE", "cs.AI", "cs.CL", "cs.MA" ], "url": "https://arxiv.org/abs/2504.20117", "pdf_url": "https://arxiv.org/pdf/2504.20117", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2505.00472", "title": "UserCentrix: An Agentic Memory-augmented AI Framework for Smart Spaces", "abstract": "Agentic Artificial Intelligence (AI) constitutes a transformative paradigm in the evolution of intelligent agents and decision-support systems, redefining smart environments by enhancing operational efficiency, optimizing resource allocation, and strengthening systemic resilience. This paper presents UserCentrix, a hybrid agentic orchestration framework for smart spaces that optimizes resource management and enhances user experience through urgency-aware and intent-driven decision-making mechanisms. The framework integrates interactive modules equipped with agentic behavior and autonomous decision-making capabilities to dynamically balance latency, accuracy, and computational cost. User intent functions as a governing control signal that prioritizes decisions, regulates task execution and resource allocation, and guides the adaptation of decision-making strategies to balance trade-offs between speed and accuracy. Experimental results demonstrate that the framework autonomously enables efficient intent processing and real-time monitoring, while balancing reasoning quality and computational efficiency, particularly under resource-constrained edge conditions.", "published": "2025-05-01", "updated": "2026-04-06", "authors": [ "Alaa Saleh", "Sasu Tarkoma", "Praveen Kumar Donta", "Anders Lindgren", "Naser Hossein Motlagh", "Schahram Dustdar", "Susanna Pirttikangas", "Lauri Lovén" ], "categories": [ "cs.AI", "cs.DC", "cs.MA", "cs.NI" ], "url": "https://arxiv.org/abs/2505.00472", "pdf_url": "https://arxiv.org/pdf/2505.00472", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2505.05177", "title": "MARK: Memory Augmented Refinement of Knowledge", "abstract": "Large Language Models (LLMs) assist in specialized tasks but struggle to align with evolving domain knowledge without costly fine-tuning. Domain knowledge consists of: Knowledge: Immutable facts (e.g., 'A stone is solid') and generally accepted principles (e.g., ethical standards); Refined Memory: Evolving insights shaped by business needs and real-world changes. However, a significant gap often exists between a domain expert's deep, nuanced understanding and the system's domain knowledge, which can hinder accurate information retrieval and application. Our Memory-Augmented Refinement of Knowledge (MARK) framework enables LLMs to continuously learn without retraining by leveraging structured refined memory, inspired by the Society of Mind. MARK operates through specialized agents, each serving a distinct role: Residual Refined Memory Agent: Stores and retrieves domain-specific insights to maintain context over time; User Question Refined Memory Agent: Captures user-provided facts, abbreviations, and terminology for better comprehension; LLM Response Refined Memory Agent: Extracts key elements from responses for refinement and personalization. These agents analyse stored refined memory, detect patterns, resolve contradictions, and improve response accuracy. Temporal factors like recency and frequency prioritize relevant information while discarding outdated insights. MARK enhances LLMs in multiple ways: Ground Truth Strategy: Reduces hallucinations by establishing a structured reference; Domain-Specific Adaptation: Essential for fields like healthcare, law, and manufacturing, where proprietary insights are absent from public datasets; Personalized AI Assistants: Improves virtual assistants by remembering user preferences, ensuring coherent responses over time.", "published": "2025-05-08", "updated": "2025-05-08", "authors": [ "Anish Ganguli", "Prabal Deb", "Debleena Banerjee" ], "categories": [ "cs.AI", "cs.LG" ], "url": "https://arxiv.org/abs/2505.05177", "pdf_url": "https://arxiv.org/pdf/2505.05177", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2505.12923", "title": "The Traitors: Deception and Trust in Multi-Agent Language Model Simulations", "abstract": "As AI systems increasingly assume roles where trust and alignment with human values are essential, understanding when and why they engage in deception has become a critical research priority. We introduce The Traitors, a multi-agent simulation framework inspired by social deduction games, designed to probe deception, trust formation, and strategic communication among large language model (LLM) agents under asymmetric information. A minority of agents the traitors seek to mislead the majority, while the faithful must infer hidden identities through dialogue and reasoning. Our contributions are: (1) we ground the environment in formal frameworks from game theory, behavioral economics, and social cognition; (2) we develop a suite of evaluation metrics capturing deception success, trust dynamics, and collective inference quality; (3) we implement a fully autonomous simulation platform where LLMs reason over persistent memory and evolving social dynamics, with support for heterogeneous agent populations, specialized traits, and adaptive behaviors. Our initial experiments across DeepSeek-V3, GPT-4o-mini, and GPT-4o (10 runs per model) reveal a notable asymmetry: advanced models like GPT-4o demonstrate superior deceptive capabilities yet exhibit disproportionate vulnerability to others' falsehoods. This suggests deception skills may scale faster than detection abilities. Overall, The Traitors provides a focused, configurable testbed for investigating LLM behavior in socially nuanced interactions. We position this work as a contribution toward more rigorous research on deception mechanisms, alignment challenges, and the broader social reliability of AI systems.", "published": "2025-05-19", "updated": "2025-12-14", "authors": [ "Pedro M. P. Curvo" ], "categories": [ "cs.AI", "cs.MA" ], "url": "https://arxiv.org/abs/2505.12923", "pdf_url": "https://arxiv.org/pdf/2505.12923", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2505.13941", "title": "MLZero: A Multi-Agent System for End-to-end Machine Learning Automation", "abstract": "Existing AutoML systems have advanced the automation of machine learning (ML); however, they still require substantial manual configuration and expert input, particularly when handling multimodal data. We introduce MLZero, a novel multi-agent framework powered by Large Language Models (LLMs) that enables end-to-end ML automation across diverse data modalities with minimal human intervention. A cognitive perception module is first employed, transforming raw multimodal inputs into perceptual context that effectively guides the subsequent workflow. To address key limitations of LLMs, such as hallucinated code generation and outdated API knowledge, we enhance the iterative code generation process with semantic and episodic memory. MLZero demonstrates superior performance on MLE-Bench Lite, outperforming all competitors in both success rate and solution quality, securing six gold medals. Additionally, when evaluated on our Multimodal AutoML Agent Benchmark, which includes 25 more challenging tasks spanning diverse data modalities, MLZero outperforms the competing methods by a large margin with a success rate of 0.92 (+263.6\\%) and an average rank of 2.28. Our approach maintains its robust effectiveness even with a compact 8B LLM, outperforming full-size systems from existing solutions.", "published": "2025-05-20", "updated": "2025-05-20", "authors": [ "Haoyang Fang", "Boran Han", "Nick Erickson", "Xiyuan Zhang", "Su Zhou", "Anirudh Dagar", "Jiani Zhang", "Ali Caner Turkmen", "Cuixiong Hu", "Huzefa Rangwala", "Ying Nian Wu", "Bernie Wang", "George Karypis" ], "categories": [ "cs.MA", "cs.AI", "cs.CL", "cs.LG" ], "url": "https://arxiv.org/abs/2505.13941", "pdf_url": "https://arxiv.org/pdf/2505.13941", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2505.14163", "title": "DSMentor: Enhancing Data Science Agents with Curriculum Learning and Online Knowledge Accumulation", "abstract": "Large language model (LLM) agents have shown promising performance in generating code for solving complex data science problems. Recent studies primarily focus on enhancing in-context learning through improved search, sampling, and planning techniques, while overlooking the importance of the order in which problems are tackled during inference. In this work, we develop a novel inference-time optimization framework, referred to as DSMentor, which leverages curriculum learning -- a strategy that introduces simpler task first and progressively moves to more complex ones as the learner improves -- to enhance LLM agent performance in challenging data science tasks. Our mentor-guided framework organizes data science tasks in order of increasing difficulty and incorporates a growing long-term memory to retain prior experiences, guiding the agent's learning progression and enabling more effective utilization of accumulated knowledge. We evaluate DSMentor through extensive experiments on DSEval and QRData benchmarks. Experiments show that DSMentor using Claude-3.5-Sonnet improves the pass rate by up to 5.2% on DSEval and QRData compared to baseline agents. Furthermore, DSMentor demonstrates stronger causal reasoning ability, improving the pass rate by 8.8% on the causality problems compared to GPT-4 using Program-of-Thoughts prompts. Our work underscores the importance of developing effective strategies for accumulating and utilizing knowledge during inference, mirroring the human learning process and opening new avenues for improving LLM performance through curriculum-based inference optimization.", "published": "2025-05-20", "updated": "2025-05-20", "authors": [ "He Wang", "Alexander Hanbo Li", "Yiqun Hu", "Sheng Zhang", "Hideo Kobayashi", "Jiani Zhang", "Henry Zhu", "Chung-Wei Hang", "Patrick Ng" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2505.14163", "pdf_url": "https://arxiv.org/pdf/2505.14163", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2505.16067", "title": "How Memory Management Impacts LLM Agents: An Empirical Study of Experience-Following Behavior", "abstract": "Memory is a critical component in large language model (LLM)-based agents, enabling them to store and retrieve past executions to improve task performance over time. In this paper, we conduct an empirical study on how memory management choices impact the LLM agents' behavior, especially their long-term performance. Specifically, we focus on two fundamental memory management operations that are widely used by many agent frameworks-memory addition and deletion-to systematically study their impact on the agent behavior. Through our quantitative analysis, we find that LLM agents display an experience-following property: high similarity between a task input and the input in a retrieved memory record often results in highly similar agent outputs. Our analysis further reveals two significant challenges associated with this property: error propagation, where inaccuracies in past experiences compound and degrade future performance, and misaligned experience replay, where some seemingly correct executions can provide limited or even misleading value as experiences. Through controlled experiments, we demonstrate the importance of regulating experience quality within the memory bank and show that future task evaluations can serve as free quality labels for stored memory. Our findings offer insights into the behavioral dynamics of LLM agent memory systems and provide practical guidance for designing memory components that support robust, long-term agent performance.", "published": "2025-05-21", "updated": "2025-10-10", "authors": [ "Zidi Xiong", "Yuping Lin", "Wenya Xie", "Pengfei He", "Zirui Liu", "Jiliang Tang", "Himabindu Lakkaraju", "Zhen Xiang" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2505.16067", "pdf_url": "https://arxiv.org/pdf/2505.16067", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2505.16348", "title": "Embodied Agents Meet Personalization: Investigating Challenges and Solutions Through the Lens of Memory Utilization", "abstract": "LLM-powered embodied agents have shown success on conventional object-rearrangement tasks, but providing personalized assistance that leverages user-specific knowledge from past interactions presents new challenges. We investigate these challenges through the lens of agents' memory utilization along two critical dimensions: object semantics (identifying objects based on personal meaning) and user patterns (recalling sequences from behavioral routines). To assess these capabilities, we construct MEMENTO, an end-to-end two-stage evaluation framework comprising single-memory and joint-memory tasks. Our experiments reveal that current agents can recall simple object semantics but struggle to apply sequential user patterns to planning. Through in-depth analysis, we identify two critical bottlenecks: information overload and coordination failures when handling multiple memories. Based on these findings, we explore memory architectural approaches to address these challenges. Given our observation that episodic memory provides both personalized knowledge and in-context learning benefits, we design a hierarchical knowledge graph-based user-profile memory module that separately manages personalized knowledge, achieving substantial improvements on both single and joint-memory tasks. Project website: https://connoriginal.github.io/MEMENTO", "published": "2025-05-22", "updated": "2026-02-13", "authors": [ "Taeyoon Kwon", "Dongwook Choi", "Hyojun Kim", "Sunghwan Kim", "Seungjun Moon", "Beong-woo Kwak", "Kuan-Hao Huang", "Jinyoung Yeo" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2505.16348", "pdf_url": "https://arxiv.org/pdf/2505.16348", "matched_queries": [ "agent-memory", "episodic-memory" ] }, { "arxiv_id": "2505.18279", "title": "Collaborative Memory: Multi-User Memory Sharing in LLM Agents with Dynamic Access Control", "abstract": "Complex tasks are increasingly delegated to ensembles of specialized LLM-based agents that reason, communicate, and coordinate actions-both among themselves and through interactions with external tools, APIs, and databases. While persistent memory has been shown to enhance single-agent performance, most approaches assume a monolithic, single-user context-overlooking the benefits and challenges of knowledge transfer across users under dynamic, asymmetric permissions. We introduce Collaborative Memory, a framework for multi-user, multi-agent environments with asymmetric, time-evolving access controls encoded as bipartite graphs linking users, agents, and resources. Our system maintains two memory tiers: (1) private memory-private fragments visible only to their originating user; and (2) shared memory-selectively shared fragments. Each fragment carries immutable provenance attributes (contributing agents, accessed resources, and timestamps) to support retrospective permission checks. Granular read policies enforce current user-agent-resource constraints and project existing memory fragments into filtered transformed views. Write policies determine fragment retention and sharing, applying context-aware transformations to update the memory. Both policies may be designed conditioned on system, agent, and user-level information. Our framework enables safe, efficient, and interpretable cross-user knowledge sharing, with provable adherence to asymmetric, time-varying policies and full auditability of memory operations.", "published": "2025-05-23", "updated": "2025-05-23", "authors": [ "Alireza Rezazadeh", "Zichao Li", "Ange Lou", "Yuying Zhao", "Wei Wei", "Yujia Bao" ], "categories": [ "cs.MA", "cs.AI", "cs.CL", "cs.LG" ], "url": "https://arxiv.org/abs/2505.18279", "pdf_url": "https://arxiv.org/pdf/2505.18279", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2505.19237", "title": "Sensorimotor Self-Recognition in Multimodal Large Language Model-Driven Robots", "abstract": "Self-recognition -- the ability to maintain an internal representation of one's own body within the environment -- underpins intelligent, autonomous behavior. As a foundational component of the minimal self, self-recognition provides the initial substrate from which higher forms of self-awareness may eventually emerge. Recent advances in large language models achieve human-like performance in tasks integrating multimodal information, raising growing interest in the embodiment capabilities of AI agents deployed on nonhuman platforms such as robots. We investigate whether multimodal LLMs can develop self-recognition through sensorimotor experience by integrating an LLM into an autonomous mobile robot. The system exhibits robust environmental awareness, self-identification, and predictive awareness, enabling it to infer its robotic nature and motion characteristics. Structural equation modeling reveals how sensory integration influences distinct dimensions of the minimal self and their coordination with past-present memory, as well as the hierarchical internal associations that drive self-identification. Ablation tests of sensory inputs demonstrate compensatory interactions among sensors and confirm the essential role of structured and episodic memory. Given appropriate sensory information about the world and itself, multimodal LLMs open the door to artificial selfhood in embodied cognitive systems.", "published": "2025-05-25", "updated": "2026-04-19", "authors": [ "Iñaki Dellibarda Varela", "Pablo Romero-Sorozabal", "Diego Torricelli", "Gabriel Delgado-Oleas", "Jose Ignacio Serrano", "Maria Dolores del Castillo Sobrino", "Eduardo Rocon", "Manuel Cebrian" ], "categories": [ "cs.AI", "cs.RO" ], "url": "https://arxiv.org/abs/2505.19237", "pdf_url": "https://arxiv.org/pdf/2505.19237", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2505.19436", "title": "Task Memory Engine: Spatial Memory for Robust Multi-Step LLM Agents", "abstract": "Large Language Models (LLMs) falter in multi-step interactions -- often hallucinating, repeating actions, or misinterpreting user corrections -- due to reliance on linear, unstructured context. This fragility stems from the lack of persistent memory to track evolving goals and task dependencies, undermining trust in autonomous agents. We introduce the Task Memory Engine (TME), a modular memory controller that transforms existing LLMs into robust, revision-aware agents without fine-tuning. TME implements a spatial memory framework that replaces flat context with graph-based structures to support consistent, multi-turn reasoning. Departing from linear concatenation and ReAct-style prompting, TME builds a dynamic task graph -- either a tree or directed acyclic graph (DAG) -- to map user inputs to subtasks, align them with prior context, and enable dependency-tracked revisions. Its Task Representation and Intent Management (TRIM) component models task semantics and user intent to ensure accurate interpretation. Across four multi-turn scenarios-trip planning, cooking, meeting scheduling, and shopping cart editing -- TME eliminates 100% of hallucinations and misinterpretations in three tasks, and reduces hallucinations by 66.7% and misinterpretations by 83.3% across 27 user turns, outperforming ReAct. TME's modular design supports plug-and-play deployment and domain-specific customization, adaptable to both personal assistants and enterprise automation. We release TME's codebase, benchmarks, and components as open-source resources, enabling researchers to develop reliable LLM agents. TME's scalable architecture addresses a critical gap in agent performance across complex, interactive settings.", "published": "2025-05-26", "updated": "2025-05-26", "authors": [ "Ye Ye" ], "categories": [ "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2505.19436", "pdf_url": "https://arxiv.org/pdf/2505.19436", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2505.20231", "title": "MemGuide: Intent-Driven Memory Selection for Goal-Oriented Multi-Session LLM Agents", "abstract": "Modern task-oriented dialogue (TOD) systems increasingly rely on large language model (LLM) agents, leveraging Retrieval-Augmented Generation (RAG) and long-context capabilities for long-term memory utilization. However, these methods are primarily based on semantic similarity, overlooking task intent and reducing task coherence in multi-session dialogues. To address this challenge, we introduce MemGuide, a two-stage framework for intent-driven memory selection. (1) Intent-Aligned Retrieval matches the current dialogue context with stored intent descriptions in the memory bank, retrieving QA-formatted memory units that share the same goal. (2) Missing-Slot Guided Filtering employs a chain-of-thought slot reasoner to enumerate unfilled slots, then uses a fine-tuned LLaMA-8B filter to re-rank the retrieved units by marginal slot-completion gain. The resulting memory units inform a proactive strategy that minimizes conversational turns by directly addressing information gaps. Based on this framework, we introduce the MS-TOD, the first multi-session TOD benchmark comprising 132 diverse personas, 956 task goals, and annotated intent-aligned memory targets, supporting efficient multi-session task completion. Evaluations on MS-TOD show that MemGuide raises the task success rate by 11% (88% -> 99%) and reduces dialogue length by 2.84 turns in multi-session settings, while maintaining parity with single-session benchmarks.", "published": "2025-05-26", "updated": "2025-08-13", "authors": [ "Yiming Du", "Bingbing Wang", "Yang He", "Bin Liang", "Baojun Wang", "Zhongyang Li", "Lin Gui", "Jeff Z. Pan", "Ruifeng Xu", "Kam-Fai Wong" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2505.20231", "pdf_url": "https://arxiv.org/pdf/2505.20231", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2505.22657", "title": "3DLLM-Mem: Long-Term Spatial-Temporal Memory for Embodied 3D Large Language Model", "abstract": "Humans excel at performing complex tasks by leveraging long-term memory across temporal and spatial experiences. In contrast, current Large Language Models (LLMs) struggle to effectively plan and act in dynamic, multi-room 3D environments. We posit that part of this limitation is due to the lack of proper 3D spatial-temporal memory modeling in LLMs. To address this, we first introduce 3DMem-Bench, a comprehensive benchmark comprising over 26,000 trajectories and 2,892 embodied tasks, question-answering and captioning, designed to evaluate an agent's ability to reason over long-term memory in 3D environments. Second, we propose 3DLLM-Mem, a novel dynamic memory management and fusion model for embodied spatial-temporal reasoning and actions in LLMs. Our model uses working memory tokens, which represents current observations, as queries to selectively attend to and fuse the most useful spatial and temporal features from episodic memory, which stores past observations and interactions. Our approach allows the agent to focus on task-relevant information while maintaining memory efficiency in complex, long-horizon environments. Experimental results demonstrate that 3DLLM-Mem achieves state-of-the-art performance across various tasks, outperforming the strongest baselines by 16.5% in success rate on 3DMem-Bench's most challenging in-the-wild embodied tasks.", "published": "2025-05-28", "updated": "2025-12-17", "authors": [ "Wenbo Hu", "Yining Hong", "Yanjun Wang", "Leison Gao", "Zibu Wei", "Xingcheng Yao", "Nanyun Peng", "Yonatan Bitton", "Idan Szpektor", "Kai-Wei Chang" ], "categories": [ "cs.CV", "cs.AI", "cs.CL", "cs.LG" ], "url": "https://arxiv.org/abs/2505.22657", "pdf_url": "https://arxiv.org/pdf/2505.22657", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2506.01442", "title": "Agentic Episodic Control", "abstract": "Reinforcement learning (RL) remains fundamentally limited by poor data efficiency and weak generalization. Prior episodic RL methods attempt to alleviate this via external memory modules, yet they suffer from two key limitations: a representation bottleneck caused by shallow encoders, and a retrieval dilemma where episodic memory is accessed indiscriminately. To address these challenges, we propose Agentic Episodic Control (AEC), a novel architecture that integrates large language models (LLMs) into episodic RL. AEC uses an LLM-based semantic augmenter to generate semantic representations from raw observations, and a critical state recognizer to selectively retrieve valuable experiences. This transforms memory usage from passive similarity matching into strategic, context-aware recall. Across five BabyAI-Text environments, AEC achieves 2-6x higher data efficiency than baselines and is the only method to solve complex tasks like UnlockLocal with over 90% success. It further demonstrates strong cross-task and cross-environment generalization, maintaining performance even under distribution shifts. AEC shows that combining LLM-derived priors with reinforcement learning yields more sample-efficient and adaptable agents.", "published": "2025-06-02", "updated": "2026-06-26", "authors": [ "Xidong Yang", "Wenhao Li", "Junjie Sheng", "Yun Hua", "Haosheng Chen", "Chuyun Shen", "Xiangfeng Wang" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2506.01442", "pdf_url": "https://arxiv.org/pdf/2506.01442", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2506.01936", "title": "Should Decision-Makers Reveal Classifiers in Online Strategic Classification?", "abstract": "Strategic classification addresses a learning problem where a decision-maker implements a classifier over agents who may manipulate their features in order to receive favorable predictions. In the standard model of online strategic classification, in each round, the decision-maker implements and publicly reveals a classifier, after which agents perfectly best respond based on this knowledge. However, in practice, whether to disclose the classifier is often debated -- some decision-makers believe that hiding the classifier can prevent misclassification errors caused by manipulation. In this paper, we formally examine how limiting the agents' access to the current classifier affects the decision-maker's performance. Specifically, we consider an extended online strategic classification setting where agents lack direct knowledge about the current classifier and instead manipulate based on a weighted average of historically implemented classifiers. Our main result shows that in this setting, the decision-maker incurs $(1-γ)^{-1}$ or $k_{\\text{in}}$ times more mistakes compared to the full-knowledge setting, where $k_{\\text{in}}$ is the maximum in-degree of the manipulation graph (representing how many distinct feature vectors can be manipulated to appear as a single one), and $γ$ is the discount factor indicating agents' memory of past classifiers. Our results demonstrate how withholding access to the classifier can backfire and degrade the decision-maker's performance in online strategic classification.", "published": "2025-06-02", "updated": "2025-06-02", "authors": [ "Han Shao", "Shuo Xie", "Kunhe Yang" ], "categories": [ "cs.GT", "cs.LG" ], "url": "https://arxiv.org/abs/2506.01936", "pdf_url": "https://arxiv.org/pdf/2506.01936", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2506.07398", "title": "G-Memory: Tracing Hierarchical Memory for Multi-Agent Systems", "abstract": "Large language model (LLM)-powered multi-agent systems (MAS) have demonstrated cognitive and execution capabilities that far exceed those of single LLM agents, yet their capacity for self-evolution remains hampered by underdeveloped memory architectures. Upon close inspection, we are alarmed to discover that prevailing MAS memory mechanisms (1) are overly simplistic, completely disregarding the nuanced inter-agent collaboration trajectories, and (2) lack cross-trial and agent-specific customization, in stark contrast to the expressive memory developed for single agents. To bridge this gap, we introduce G-Memory, a hierarchical, agentic memory system for MAS inspired by organizational memory theory, which manages the lengthy MAS interaction via a three-tier graph hierarchy: insight, query, and interaction graphs. Upon receiving a new user query, G-Memory performs bi-directional memory traversal to retrieve both $\\textit{high-level, generalizable insights}$ that enable the system to leverage cross-trial knowledge, and $\\textit{fine-grained, condensed interaction trajectories}$ that compactly encode prior collaboration experiences. Upon task execution, the entire hierarchy evolves by assimilating new collaborative trajectories, nurturing the progressive evolution of agent teams. Extensive experiments across five benchmarks, three LLM backbones, and three popular MAS frameworks demonstrate that G-Memory improves success rates in embodied action and accuracy in knowledge QA by up to $20.89\\%$ and $10.12\\%$, respectively, without any modifications to the original frameworks. Our codes are available at https://github.com/bingreeky/GMemory.", "published": "2025-06-09", "updated": "2025-06-16", "authors": [ "Guibin Zhang", "Muxin Fu", "Guancheng Wan", "Miao Yu", "Kun Wang", "Shuicheng Yan" ], "categories": [ "cs.MA", "cs.CL", "cs.LG" ], "url": "https://arxiv.org/abs/2506.07398", "pdf_url": "https://arxiv.org/pdf/2506.07398", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2506.12088", "title": "Risks & Benefits of LLMs & GenAI for Platform Integrity, Healthcare Diagnostics, Financial Trust and Compliance, Cybersecurity, Privacy & AI Safety: A Comprehensive Survey, Roadmap & Implementation Blueprint", "abstract": "Large Language Models (LLMs) and generative AI (GenAI) systems, such as ChatGPT, Claude, Gemini, LLaMA, and Copilot (by OpenAI, Anthropic, Google, Meta, and Microsoft, respectively), are reshaping digital platforms and app ecosystems while introducing critical challenges in cybersecurity, privacy, and platform integrity. Our analysis reveals alarming trends: LLM-assisted malware is projected to rise from 2% (2021) to 50% (2025); AI-generated Google reviews grew nearly tenfold (1.2% in 2021 to 12.21% in 2023, expected to reach 30% by 2025); AI scam reports surged 456%; misinformation sites increased over 1500%; and deepfake attacks are projected to rise over 900% in 2025. In finance, LLM-driven threats like synthetic identity fraud and AI-generated scams are accelerating. Platforms such as JPMorgan Chase, Stripe, and Plaid deploy LLMs for fraud detection, regulation parsing, and KYC/AML automation, reducing fraud loss by up to 21% and accelerating onboarding by 40-60%. LLM-facilitated code development has driven mobile app submissions from 1.8 million (2020) to 3.0 million (2024), projected to reach 3.6 million (2025). To address AI threats, platforms like Google Play, Apple App Store, GitHub Copilot, TikTok, Facebook, and Amazon deploy LLM-based defenses, highlighting their dual nature as both threat sources and mitigation tools. In clinical diagnostics, LLMs raise concerns about accuracy, bias, and safety, necessitating strong governance. Drawing on 445 references, this paper surveys LLM/GenAI and proposes a strategic roadmap and operational blueprint integrating policy auditing (such as CCPA and GDPR compliance), fraud detection, and demonstrates an advanced LLM-DA stack with modular components, multi-LLM routing, agentic memory, and governance layers. We provide actionable insights, best practices, and real-world case studies for scalable trust and responsible innovation.", "published": "2025-06-10", "updated": "2025-07-26", "authors": [ "Kiarash Ahi" ], "categories": [ "cs.CR", "cs.CY" ], "url": "https://arxiv.org/abs/2506.12088", "pdf_url": "https://arxiv.org/pdf/2506.12088", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2506.17318", "title": "Context manipulation attacks : Web agents are susceptible to corrupted memory", "abstract": "Autonomous web navigation agents, which translate natural language instructions into sequences of browser actions, are increasingly deployed for complex tasks across e-commerce, information retrieval, and content discovery. Due to the stateless nature of large language models (LLMs), these agents rely heavily on external memory systems to maintain context across interactions. Unlike centralized systems where context is securely stored server-side, agent memory is often managed client-side or by third-party applications, creating significant security vulnerabilities. This was recently exploited to attack production systems. We introduce and formalize \"plan injection,\" a novel context manipulation attack that corrupts these agents' internal task representations by targeting this vulnerable context. Through systematic evaluation of two popular web agents, Browser-use and Agent-E, we show that plan injections bypass robust prompt injection defenses, achieving up to 3x higher attack success rates than comparable prompt-based attacks. Furthermore, \"context-chained injections,\" which craft logical bridges between legitimate user goals and attacker objectives, lead to a 17.7% increase in success rate for privacy exfiltration tasks. Our findings highlight that secure memory handling must be a first-class concern in agentic systems.", "published": "2025-06-18", "updated": "2025-06-18", "authors": [ "Atharv Singh Patlan", "Ashwin Hebbar", "Pramod Viswanath", "Prateek Mittal" ], "categories": [ "cs.CR", "cs.AI" ], "url": "https://arxiv.org/abs/2506.17318", "pdf_url": "https://arxiv.org/pdf/2506.17318", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2506.17338", "title": "PBFT-Backed Semantic Voting for Multi-Agent Memory Pruning", "abstract": "The proliferation of multi-agent systems (MAS) in complex, dynamic environments necessitates robust and efficient mechanisms for managing shared knowledge. A critical challenge is ensuring that distributed memories remain synchronized, relevant, and free from the accumulation of outdated or inconsequential data - a process analogous to biological forgetting. This paper introduces the Co-Forgetting Protocol, a novel, comprehensive framework designed to address this challenge by enabling synchronized memory pruning in MAS. The protocol integrates three key components: (1) context-aware semantic voting, where agents utilize a lightweight DistilBERT model to assess the relevance of memory items based on their content and the current operational context; (2) multi-scale temporal decay functions, which assign diminishing importance to memories based on their age and access frequency across different time horizons; and (3) a Practical Byzantine Fault Tolerance (PBFT)-based consensus mechanism, ensuring that decisions to retain or discard memory items are agreed upon by a qualified and fault-tolerant majority of agents, even in the presence of up to f Byzantine (malicious or faulty) agents in a system of N greater than or equal to 3f+1 agents. The protocol leverages gRPC for efficient inter-agent communication and Pinecone for scalable vector embedding storage and similarity search, with SQLite managing metadata. Experimental evaluations in a simulated MAS environment with four agents demonstrate the protocol's efficacy, achieving a 52% reduction in memory footprint over 500 epochs, 88% voting accuracy in forgetting decisions against human-annotated benchmarks, a 92% PBFT consensus success rate under simulated Byzantine conditions, and an 82% cache hit rate for memory access.", "published": "2025-06-19", "updated": "2025-06-24", "authors": [ "Duong Bach" ], "categories": [ "cs.DC", "cs.AI", "cs.MA" ], "url": "https://arxiv.org/abs/2506.17338", "pdf_url": "https://arxiv.org/pdf/2506.17338", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2506.19433", "title": "Mem4Nav: Boosting Vision-and-Language Navigation in Urban Environments with a Hierarchical Spatial-Cognition Long-Short Memory System", "abstract": "Vision-and-Language Navigation (VLN) in large-scale urban environments requires embodied agents to ground linguistic instructions in complex scenes and recall relevant experiences over extended time horizons. Prior modular pipelines offer interpretability but lack unified memory, while end-to-end (M)LLM agents excel at fusing vision and language yet remain constrained by fixed context windows and implicit spatial reasoning. We introduce \\textbf{Mem4Nav}, a hierarchical spatial-cognition long-short memory system that can augment any VLN backbone. Mem4Nav fuses a sparse octree for fine-grained voxel indexing with a semantic topology graph for high-level landmark connectivity, storing both in trainable memory tokens embedded via a reversible Transformer. Long-term memory (LTM) compresses and retains historical observations at both octree and graph nodes, while short-term memory (STM) caches recent multimodal entries in relative coordinates for real-time obstacle avoidance and local planning. At each step, STM retrieval sharply prunes dynamic context, and, when deeper history is needed, LTM tokens are decoded losslessly to reconstruct past embeddings. Evaluated on Touchdown and Map2Seq across three backbones (modular, state-of-the-art VLN with prompt-based LLM, and state-of-the-art VLN with strided-attention MLLM), Mem4Nav yields 7-13 pp gains in Task Completion, sufficient SPD reduction, and >10 pp nDTW improvement. Ablations confirm the indispensability of both the hierarchical map and dual memory modules. Our codes are open-sourced via https://github.com/tsinghua-fib-lab/Mem4Nav.", "published": "2025-06-24", "updated": "2025-10-10", "authors": [ "Lixuan He", "Haoyu Dong", "Zhenxing Chen", "Yangcheng Yu", "Jie Feng", "Yong Li" ], "categories": [ "cs.CV", "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2506.19433", "pdf_url": "https://arxiv.org/pdf/2506.19433", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2507.02259", "title": "MemAgent: Reshaping Long-Context LLM with Multi-Conv RL-based Memory Agent", "abstract": "Despite improvements by length extrapolation, efficient attention and memory modules, handling infinitely long documents with linear complexity without performance degradation during extrapolation remains the ultimate challenge in long-text processing. We directly optimize for long-text tasks in an end-to-end fashion and introduce a novel agent workflow, MemAgent, which reads text in segments and updates the memory using an overwrite strategy. We extend the DAPO algorithm to facilitate training via independent-context multi-conversation generation. MemAgent has demonstrated superb long-context capabilities, being able to extrapolate from an 8K context trained on 32K text to a 3.5M QA task with performance loss < 5% and achieves 95%+ in 512K RULER test.", "published": "2025-07-03", "updated": "2025-07-03", "authors": [ "Hongli Yu", "Tinghong Chen", "Jiangtao Feng", "Jiangjie Chen", "Weinan Dai", "Qiying Yu", "Ya-Qin Zhang", "Wei-Ying Ma", "Jingjing Liu", "Mingxuan Wang", "Hao Zhou" ], "categories": [ "cs.CL", "cs.AI", "cs.LG" ], "url": "https://arxiv.org/abs/2507.02259", "pdf_url": "https://arxiv.org/pdf/2507.02259", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2507.04172", "title": "Gathering Teams of Bounded Memory Agents on a Line", "abstract": "Several mobile agents, modelled as deterministic automata, navigate in an infinite line in synchronous rounds. All agents start in the same round. In each round, an agent can move to one of the two neighboring nodes, or stay idle. Agents have distinct labels which are integers from the set $\\{1,\\dots, L\\}$. They start in teams, and all agents in a team have the same starting node. The adversary decides the compositions of teams, and their starting nodes. Whenever an agent enters a node, it sees the entry port number and the states of all collocated agents; this information forms the input of the agent on the basis of which it transits to the next state and decides the current action. The aim is for all agents to gather at the same node and stop. Gathering is feasible, if this task can be accomplished for any decisions of the adversary, and its time is the worst-case number of rounds from the start till gathering. We consider the feasibility and time complexity of gathering teams of agents, and give a complete solution of this problem. It turns out that both feasibility and complexity of gathering depend on the sizes of teams. We first concentrate on the case when all teams have the same size $x$. For the oriented line, gathering is impossible if $x=1$, and it can be accomplished in time $O(D)$, for $x>1$, where $D$ is the distance between the starting nodes of the most distant teams. This complexity is of course optimal. For the unoriented line, the situation is different. For $x=1$, gathering is also impossible, but for $x=2$, the optimal time of gathering is $Θ(D\\log L)$, and for $x\\geq 3$, the optimal time of gathering is $Θ(D)$. In the case when there are teams of different sizes, we show that gathering is always possible in time $O(D)$, even for the unoriented line. This complexity is of course optimal.", "published": "2025-07-05", "updated": "2025-07-05", "authors": [ "Younan Gao", "Andrzej Pelc" ], "categories": [ "cs.DC" ], "url": "https://arxiv.org/abs/2507.04172", "pdf_url": "https://arxiv.org/pdf/2507.04172", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2507.10562", "title": "SAMEP: A Secure Protocol for Persistent Context Sharing Across AI Agents", "abstract": "Current AI agent architectures suffer from ephemeral memory limitations, preventing effective collaboration and knowledge sharing across sessions and agent boundaries. We introduce SAMEP (Secure Agent Memory Exchange Protocol), a novel framework that enables persistent, secure, and semantically searchable memory sharing among AI agents. Our protocol addresses three critical challenges: (1) persistent context preservation across agent sessions, (2) secure multi-agent collaboration with fine-grained access control, and (3) efficient semantic discovery of relevant historical context. SAMEP implements a distributed memory repository with vector-based semantic search, cryptographic access controls (AES-256-GCM), and standardized APIs compatible with existing agent communication protocols (MCP, A2A). We demonstrate SAMEP's effectiveness across diverse domains including multi-agent software development, healthcare AI with HIPAA compliance, and multi-modal processing pipelines. Experimental results show 73% reduction in redundant computations, 89% improvement in context relevance scores, and complete compliance with regulatory requirements including audit trail generation. SAMEP enables a new paradigm of persistent, collaborative AI agent ecosystems while maintaining security and privacy guarantees.", "published": "2025-07-05", "updated": "2025-07-05", "authors": [ "Hari Masoor" ], "categories": [ "cs.AI", "cs.CR", "cs.DB", "cs.LG" ], "url": "https://arxiv.org/abs/2507.10562", "pdf_url": "https://arxiv.org/pdf/2507.10562", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2507.05257", "title": "Evaluating Memory in LLM Agents via Incremental Multi-Turn Interactions", "abstract": "Recent benchmarks for Large Language Model (LLM) agents primarily focus on evaluating reasoning, planning, and execution capabilities, while another critical component-memory, encompassing how agents memorize, update, and retrieve long-term information-is under-evaluated due to the lack of benchmarks. We term agents with memory mechanisms as memory agents. In this paper, based on classic theories from memory science and cognitive science, we identify four core competencies essential for memory agents: accurate retrieval, test-time learning, long-range understanding, and selective forgetting. Existing benchmarks either rely on limited context lengths or are tailored for static, long-context settings like book-based QA, which do not reflect the interactive, multi-turn nature of memory agents that incrementally accumulate information. Moreover, no existing benchmarks cover all four competencies. We introduce MemoryAgentBench, a new benchmark specifically designed for memory agents. Our benchmark transforms existing long-context datasets and incorporates newly constructed datasets into a multi-turn format, effectively simulating the incremental information processing characteristic of memory agents. By carefully selecting and curating datasets, our benchmark provides comprehensive coverage of the four core memory competencies outlined above, thereby offering a systematic and challenging testbed for assessing memory quality. We evaluate a diverse set of memory agents, ranging from simple context-based and retrieval-augmented generation (RAG) systems to advanced agents with external memory modules and tool integration. Empirical results reveal that current methods fall short of mastering all four competencies, underscoring the need for further research into comprehensive memory mechanisms for LLM agents.", "published": "2025-07-07", "updated": "2026-06-28", "authors": [ "Yuanzhe Hu", "Yu Wang", "Julian McAuley" ], "categories": [ "cs.CL", "cs.AI" ], "url": "https://arxiv.org/abs/2507.05257", "pdf_url": "https://arxiv.org/pdf/2507.05257", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2507.05445", "title": "A Systematization of Security Vulnerabilities in Computer Use Agents", "abstract": "Computer Use Agents (CUAs), autonomous systems that interact with software interfaces via browsers or virtual machines, are rapidly being deployed in consumer and enterprise environments. These agents introduce novel attack surfaces and trust boundaries that are not captured by traditional threat models. Despite their growing capabilities, the security boundaries of CUAs remain poorly understood. In this paper, we conduct a systematic threat analysis and testing of real-world CUAs under adversarial conditions. We identify seven classes of risks unique to the CUA paradigm, and analyze three concrete exploit scenarios in depth: (1) clickjacking via visual overlays that mislead interface-level reasoning, (2) indirect prompt injection that enables Remote Code Execution (RCE) through chained tool use, and (3) CoT exposure attacks that manipulate implicit interface framing to hijack multi-step reasoning. These case studies reveal deeper architectural flaws across current CUA implementations. Namely, a lack of input provenance tracking, weak interface-action binding, and insufficient control over agent memory and delegation. We conclude by proposing a CUA-specific security evaluation framework and design principles for safe deployment in adversarial and high-stakes settings.", "published": "2025-07-07", "updated": "2025-07-07", "authors": [ "Daniel Jones", "Giorgio Severi", "Martin Pouliot", "Gary Lopez", "Joris de Gruyter", "Santiago Zanella-Beguelin", "Justin Song", "Blake Bullwinkel", "Pamela Cortez", "Amanda Minnich" ], "categories": [ "cs.CR" ], "url": "https://arxiv.org/abs/2507.05445", "pdf_url": "https://arxiv.org/pdf/2507.05445", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2507.07957", "title": "MIRIX: Multi-Agent Memory System for LLM-Based Agents", "abstract": "Although memory capabilities of AI agents are gaining increasing attention, existing solutions remain fundamentally limited. Most rely on flat, narrowly scoped memory components, constraining their ability to personalize, abstract, and reliably recall user-specific information over time. To this end, we introduce MIRIX, a modular, multi-agent memory system that redefines the future of AI memory by solving the field's most critical challenge: enabling language models to truly remember. Unlike prior approaches, MIRIX transcends text to embrace rich visual and multimodal experiences, making memory genuinely useful in real-world scenarios. MIRIX consists of six distinct, carefully structured memory types: Core, Episodic, Semantic, Procedural, Resource Memory, and Knowledge Vault, coupled with a multi-agent framework that dynamically controls and coordinates updates and retrieval. This design enables agents to persist, reason over, and accurately retrieve diverse, long-term user data at scale. We validate MIRIX in two demanding settings. First, on ScreenshotVQA, a challenging multimodal benchmark comprising nearly 20,000 high-resolution computer screenshots per sequence, requiring deep contextual understanding and where no existing memory systems can be applied, MIRIX achieves 35% higher accuracy than the RAG baseline while reducing storage requirements by 99.9%. Second, on LOCOMO, a long-form conversation benchmark with single-modal textual input, MIRIX attains state-of-the-art performance of 85.4%, far surpassing existing baselines. These results show that MIRIX sets a new performance standard for memory-augmented LLM agents. To allow users to experience our memory system, we provide a packaged application powered by MIRIX. It monitors the screen in real time, builds a personalized memory base, and offers intuitive visualization and secure local storage to ensure privacy.", "published": "2025-07-10", "updated": "2025-07-10", "authors": [ "Yu Wang", "Xi Chen" ], "categories": [ "cs.CL", "cs.AI" ], "url": "https://arxiv.org/abs/2507.07957", "pdf_url": "https://arxiv.org/pdf/2507.07957", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2507.16713", "title": "A Pragmatist Robot: Learning to Plan Tasks by Experiencing the Real World", "abstract": "Large language models (LLMs) have emerged as the dominant paradigm for robotic task planning using natural language instructions. However, trained on general internet data, LLMs are not inherently aligned with the embodiment, skill sets, and limitations of real-world robotic systems. Inspired by the emerging paradigm of verbal reinforcement learning-where LLM agents improve through self-reflection and few-shot learning without parameter updates-we introduce PragmaBot, a framework that enables robots to learn task planning through real-world experience. PragmaBot employs a vision-language model (VLM) as the robot's \"brain\" and \"eye\", allowing it to visually evaluate action outcomes and self-reflect on failures. These reflections are stored in a short-term memory (STM), enabling the robot to quickly adapt its behavior during ongoing tasks. Upon task completion, the robot summarizes the lessons learned into its long-term memory (LTM). When facing new tasks, it can leverage retrieval-augmented generation (RAG) to plan more grounded action sequences by drawing on relevant past experiences and knowledge. Experiments on four challenging robotic tasks show that STM-based self-reflection increases task success rates from 35% to 84%, with emergent intelligent object interactions. In 12 real-world scenarios (including eight previously unseen tasks), the robot effectively learns from the LTM and improves single-trial success rates from 22% to 80%, with RAG outperforming naive prompting. These results highlight the effectiveness and generalizability of PragmaBot. Project webpage: https://pragmabot.github.io/", "published": "2025-07-22", "updated": "2026-02-14", "authors": [ "Kaixian Qu", "Guowei Lan", "René Zurbrügg", "Changan Chen", "Christopher E. Mower", "Haitham Bou-Ammar", "Marco Hutter" ], "categories": [ "cs.RO", "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2507.16713", "pdf_url": "https://arxiv.org/pdf/2507.16713", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2507.22925", "title": "Hierarchical Memory for High-Efficiency Long-Term Reasoning in LLM Agents", "abstract": "Long-term memory is one of the key factors influencing the reasoning capabilities of Large Language Model Agents (LLM Agents). Incorporating a memory mechanism that effectively integrates past interactions can significantly enhance decision-making and contextual coherence of LLM Agents. While recent works have made progress in memory storage and retrieval, such as encoding memory into dense vectors for similarity-based search or organizing knowledge in the form of graph, these approaches often fall short in structured memory organization and efficient retrieval. To address these limitations, we propose a Hierarchical Memory (H-MEM) architecture for LLM Agents that organizes and updates memory in a multi-level fashion based on the degree of semantic abstraction. Each memory vector is embedded with a positional index encoding pointing to its semantically related sub-memories in the next layer. During the reasoning phase, an index-based routing mechanism enables efficient, layer-by-layer retrieval without performing exhaustive similarity computations. We evaluate our method on five task settings from the LoCoMo dataset. Experimental results show that our approach consistently outperforms five baseline methods, demonstrating its effectiveness in long-term dialogue scenarios.", "published": "2025-07-23", "updated": "2025-07-23", "authors": [ "Haoran Sun", "Shaoning Zeng" ], "categories": [ "cs.CL", "cs.AI" ], "url": "https://arxiv.org/abs/2507.22925", "pdf_url": "https://arxiv.org/pdf/2507.22925", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2507.21407", "title": "Graph-Augmented Large Language Model Agents: Current Progress and Future Prospects", "abstract": "Autonomous agents based on large language models (LLMs) have demonstrated impressive capabilities in a wide range of applications, including web navigation, software development, and embodied control. While most LLMs are limited in several key agentic procedures, such as reliable planning, long-term memory, tool management, and multi-agent coordination, graphs can serve as a powerful auxiliary structure to enhance structure, continuity, and coordination in complex agent workflows. Given the rapid growth and fragmentation of research on Graph-augmented LLM Agents (GLA), this paper offers a timely and comprehensive overview of recent advances and also highlights key directions for future work. Specifically, we categorize existing GLA methods by their primary functions in LLM agent systems, including planning, memory, and tool usage, and then analyze how graphs and graph learning algorithms contribute to each. For multi-agent systems, we further discuss how GLA solutions facilitate the orchestration, efficiency optimization, and trustworthiness of MAS. Finally, we highlight key future directions to advance this field, from improving structural adaptability to enabling unified, scalable, and multimodal GLA systems. We hope this paper can serve as a roadmap for future research on GLA and foster a deeper understanding of the role of graphs in LLM agent systems.", "published": "2025-07-29", "updated": "2025-08-30", "authors": [ "Yixin Liu", "Guibin Zhang", "Kun Wang", "Shiyuan Li", "Shirui Pan" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2507.21407", "pdf_url": "https://arxiv.org/pdf/2507.21407", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2508.00031", "title": "Git Context Controller: Manage the Context of LLM-based Agents like Git", "abstract": "Large language model (LLM) agents have demonstrated strong capabilities in long-horizon tasks by interleaving reasoning with tool use. However, as these agents scale to complex workflows such as software engineering and open-ended research, context management becomes a fundamental bottleneck: interaction histories grow unbounded, become costly to maintain, and are difficult to reuse across sessions and agents. We introduce \\textbf{Git-Context-Controller (GCC)}, a structured context management framework inspired by software version control systems. GCC elevates agent context from a transient token stream to a persistent, navigable memory workspace with explicit operations -- \\texttt{COMMIT}, \\texttt{BRANCH}, \\texttt{MERGE}, and \\texttt{CONTEXT}, that enable milestone-based checkpointing, isolated exploration of alternative reasoning paths, and hierarchical retrieval of historical context. By organizing agent memory as a versioned file system, GCC allows agents to manage long-term goals, recover and transfer reasoning across sessions, and coordinate multi-trajectory problem solving in a principled manner. Empirically, agents equipped with GCC achieve state-of-the-art performance on both SWE-Bench and BrowseComp benchmarks. On SWE-Bench Verified, GCC improves task resolution by over 13\\% relative to strong long-context baselines and outperforms 26 existing open and commercial systems, reaching over 80\\% success rate. The project will be open-sourced for the research community. The algorithm has been incorporated to the project: https://one-context.com", "published": "2025-07-30", "updated": "2026-03-01", "authors": [ "Junde Wu", "Minhao Hu", "Jiayuan Zhu", "Jiazhen Pan", "Yuyuan Liu", "Min Xu", "Yueming Jin" ], "categories": [ "cs.SE" ], "url": "https://arxiv.org/abs/2508.00031", "pdf_url": "https://arxiv.org/pdf/2508.00031", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2508.01287", "title": "Exploitation Is All You Need... for Exploration", "abstract": "Ensuring sufficient exploration is a central challenge when training meta-reinforcement learning (meta-RL) agents to solve novel environments. Conventional solutions to the exploration-exploitation dilemma inject explicit incentives such as randomization, uncertainty bonuses, or intrinsic rewards to encourage exploration. In this work, we hypothesize that an agent trained solely to maximize a greedy (exploitation-only) objective can nonetheless exhibit emergent exploratory behavior, provided three conditions are met: (1) Recurring Environmental Structure, where the environment features repeatable regularities that allow past experience to inform future choices; (2) Agent Memory, enabling the agent to retain and utilize historical interaction data; and (3) Long-Horizon Credit Assignment, where learning propagates returns over a time frame sufficient for the delayed benefits of exploration to inform current decisions. Through experiments in stochastic multi-armed bandits and temporally extended gridworlds, we observe that, when both structure and memory are present, a policy trained on a strictly greedy objective exhibits information-seeking exploratory behavior. We further demonstrate, through controlled ablations, that emergent exploration vanishes if either environmental structure or agent memory is absent (Conditions 1 & 2). Surprisingly, removing long-horizon credit assignment (Condition 3) does not always prevent emergent exploration-a result we attribute to the pseudo-Thompson Sampling effect. These findings suggest that, under the right prerequisites, exploration and exploitation need not be treated as orthogonal objectives but can emerge from a unified reward-maximization process.", "published": "2025-08-02", "updated": "2025-08-02", "authors": [ "Micah Rentschler", "Jesse Roberts" ], "categories": [ "cs.LG", "cs.AI" ], "url": "https://arxiv.org/abs/2508.01287", "pdf_url": "https://arxiv.org/pdf/2508.01287", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2508.01415", "title": "RoboMemory: A Brain-inspired Multi-memory Agentic Framework for Interactive Environmental Learning in Physical Embodied Systems", "abstract": "Embodied intelligence aims to enable robots to learn, reason, and generalize robustly across complex real-world environments. However, existing approaches often struggle with partial observability, fragmented spatial reasoning, and inefficient integration of heterogeneous memories, limiting their capacity for long-horizon adaptation. To address this, we introduce RoboMemory, a brain-inspired framework that unifies Spatial, Temporal, Episodic, and Semantic memory within a parallelized architecture for efficient long-horizon planning and interactive learning. Its core innovations are a dynamic spatial knowledge graph for scalable, consistent memory updates and a closed-loop planner with a critic module for adaptive decision-making. Extensive experiments on EmbodiedBench show that RoboMemory, instantiated with Qwen2.5-VL-72B-Ins, improves the average success rate by 26.5% over its strong baseline and even surpasses the closed-source SOTA, Claude-3.5-Sonnet. Real-world trials further confirm its capability for cumulative learning, with performance consistently improving over repeated tasks. Our results position RoboMemory as a scalable foundation for memory-augmented embodied agents, bridging insights from cognitive neuroscience with practical robotic autonomy.", "published": "2025-08-02", "updated": "2026-03-24", "authors": [ "Mingcong Lei", "Honghao Cai", "Yuyuan Yang", "Yimou Wu", "Jinke Ren", "Zezhou Cui", "Liangchen Tan", "Junkun Hong", "Gehan Hu", "Shuangyu Zhu", "Shaohan Jiang", "Ge Wang", "Junyuan Tan", "Zhenglin Wan", "Zheng Li", "Zhen Li", "Shuguang Cui", "Yiming Zhao", "Yatong Han" ], "categories": [ "cs.RO", "cs.AI" ], "url": "https://arxiv.org/abs/2508.01415", "pdf_url": "https://arxiv.org/pdf/2508.01415", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2508.03341", "title": "What Deserves Memory: Adaptive Memory Distillation for LLM Agents", "abstract": "Memory systems for LLM agents struggle to determine what information deserves retention. Existing approaches rely on predefined heuristics such as importance scores, emotional tags, or factual templates, encoding designer intuition rather than learning from the data itself. Inspired by cognitive ideas, we propose NEMORI, an adaptive memory distillation framework that casts the assessment of experience's future utility as a matter of predictability. Specifically, NEMORI comprises two cascading modules: Episodic Memory Integration transforms raw interactions into coherent narratives, and Semantic Knowledge Distillation extracts insights via prediction error. Centering on distillation, the framework remains agnostic to downstream management. Extensive experiments confirm that NEMORI achieves strong performance, efficiency, and storage reduction. Our work suggests that observing the intrinsic properties of interaction sequences offers a viable, data-driven alternative to heuristic-based memory design. Code: https://github.com/nemori-ai/nemori.", "published": "2025-08-05", "updated": "2026-04-16", "authors": [ "Wenquan Ma", "Jiayan Nan", "Wenlong Wu", "Yize Chen" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2508.03341", "pdf_url": "https://arxiv.org/pdf/2508.03341", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2508.05002", "title": "AgenticData: An Agentic Data Analytics System for Heterogeneous Data", "abstract": "Existing unstructured data analytics systems rely on experts to write code and manage complex analysis workflows, making them both expensive and time-consuming. To address these challenges, we introduce AgenticData, an innovative agentic data analytics system that allows users to simply pose natural language (NL) questions while autonomously analyzing data sources across multiple domains, including both unstructured and structured data. First, AgenticData employs a feedback-driven planning technique that automatically converts an NL query into a semantic plan composed of relational and semantic operators. We propose a multi-agent collaboration strategy by utilizing a data profiling agent for discovering relevant data, a semantic cross-validation agent for iterative optimization based on feedback, and a smart memory agent for maintaining short-term context and long-term knowledge. Second, we propose a semantic optimization model to refine and execute semantic plans effectively. Our system, AgenticData, has been tested using three benchmarks. Experimental results showed that AgenticData achieved superior accuracy on both easy and difficult tasks, significantly outperforming state-of-the-art methods.", "published": "2025-08-07", "updated": "2025-08-07", "authors": [ "Ji Sun", "Guoliang Li", "Peiyao Zhou", "Yihui Ma", "Jingzhe Xu", "Yuan Li" ], "categories": [ "cs.DB", "cs.AI" ], "url": "https://arxiv.org/abs/2508.05002", "pdf_url": "https://arxiv.org/pdf/2508.05002", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2508.07010", "title": "Narrative Memory in Machines: Multi-Agent Arc Extraction in Serialized TV", "abstract": "Serialized television narratives present significant analytical challenges due to their complex, temporally distributed storylines that necessitate sophisticated information management. This paper introduces a multi-agent system (MAS) designed to extract and analyze narrative arcs by implementing principles of computational memory architectures. The system conceptualizes narrative understanding through analogues of human memory: Large Language Models (LLMs) provide a form of semantic memory for general narrative patterns, while a vector database stores specific arc progressions as episodic memories. A multi-agent workflow simulates working memory processes to integrate these information types. Tested on the first season of Grey's Anatomy (ABC 2005-), the MAS identifies three arc types: Anthology (self-contained), Soap (relationship-focused), and Genre-Specific. These arcs and their episodic developments are stored in a vector database, facilitating structured analysis and semantic comparison. To bridge automation with critical interpretation, a graphical interface enables human oversight and refinement of the system's narrative memory. While demonstrating strong performance in identifying Anthology Arcs and character entities, the system's reliance on textual paratexts (episode summaries) revealed limitations in discerning overlapping arcs and opaque dynamics, underscoring the challenges in computational memory consolidation versus human holistic understanding. This memory-centric approach highlights the potential of combining AI-driven memory processing with human expertise. Beyond television, it offers promise for serialized written formats where narrative is entirely text-based. Future work will focus on integrating multimodal inputs to enrich episodic memory, refining memory integration mechanisms within the MAS, and expanding testing across diverse genres.", "published": "2025-08-09", "updated": "2025-08-09", "authors": [ "Roberto Balestri", "Guglielmo Pescatore" ], "categories": [ "cs.MM", "cs.HC", "cs.MA" ], "url": "https://arxiv.org/abs/2508.07010", "pdf_url": "https://arxiv.org/pdf/2508.07010", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2508.08997", "title": "Intrinsic Memory Agents: Heterogeneous Multi-Agent LLM Systems through Structured Contextual Memory", "abstract": "Multi-agent systems built on Large Language Models (LLMs) show exceptional promise for complex collaborative problem-solving, yet they face fundamental challenges stemming from context window limitations that impair memory consistency, role adherence, and procedural integrity. This paper introduces Intrinsic Memory Agents, a novel framework that addresses these limitations through agent-specific memories that evolve intrinsically with agent outputs. Specifically, our method maintains role-aligned memory that preserves specialized perspectives while focusing on task-relevant information. Our approach utilises a generic memory template applicable to new problems without the need to hand-craft specific memory prompts. We benchmark our approach on the PDDL, FEVER, and ALFWorld datasets, comparing its performance to existing state-of-the-art multi-agentic memory approaches and showing state-of-the-art or comparable performance across all three, with the highest consistency. An additional evaluation is performed on a complex data pipeline design task, and we demonstrate that our approach produces higher quality designs across 5 metrics: scalability, reliability, usability, cost-effectiveness, and documentation, plus additional qualitative evidence of the improvements. Our findings suggest that addressing memory limitations through intrinsic approaches can improve the capabilities of multi-agent LLM systems on structured planning tasks.", "published": "2025-08-12", "updated": "2026-01-12", "authors": [ "Sizhe Yuen", "Francisco Gomez Medina", "Ting Su", "Yali Du", "Adam J. Sobey" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2508.08997", "pdf_url": "https://arxiv.org/pdf/2508.08997", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2508.09486", "title": "Video-EM: Event-Centric Episodic Memory for Long-Form Video Understanding", "abstract": "Video Large Language Models (Video-LLMs) have shown strong video understanding, yet their application to long-form videos remains constrained by limited context windows. A common workaround is to compress long videos into a handful of representative frames via retrieval or summarization. However, most existing pipelines score frames in isolation, implicitly assuming that frame-level saliency is sufficient for downstream reasoning. This often yields redundant selections, fragmented temporal evidence, and weakened narrative grounding for long-form video question answering. We present \\textbf{Video-EM}, a training-free, event-centric episodic memory framework that reframes long-form VideoQA as \\emph{episodic event construction} followed by \\emph{memory refinement}. Instead of treating retrieved keyframes as independent visuals, Video-EM employs an LLM as an active memory agent to orchestrate off-the-shelf tools: it first localizes query-relevant moments via multi-grained semantic matching, then groups and segments them into temporally coherent events, and finally encodes each event as a grounded episodic memory with explicit temporal indices and spatio-temporal cues (capturing \\emph{when}, \\emph{where}, \\emph{what}, and involved entities). To further suppress verbosity and noise from imperfect upstream signals, Video-EM integrates a reasoning-driven self-reflection loop that iteratively verifies evidence sufficiency and cross-event consistency, removes redundancy, and adaptively adjusts event granularity. The outcome is a compact yet reliable \\emph{event timeline} -- a minimal but sufficient episodic memory set that can be directly consumed by existing Video-LLMs without additional training or architectural changes.", "published": "2025-08-13", "updated": "2026-03-07", "authors": [ "Yun Wang", "Long Zhang", "Jingren Liu", "Jiaqi Yan", "Zhanjie Zhang", "Jiahao Zheng", "Ao Ma", "Run Ling", "Xun Yang", "Dapeng Wu", "Xiangyu Chen", "Xuelong Li" ], "categories": [ "cs.CV", "cs.AI", "cs.MM" ], "url": "https://arxiv.org/abs/2508.09486", "pdf_url": "https://arxiv.org/pdf/2508.09486", "matched_queries": [ "agent-memory", "episodic-memory" ] }, { "arxiv_id": "2508.12630", "title": "Semantic Anchoring in Agentic Memory: Leveraging Linguistic Structures for Persistent Conversational Context", "abstract": "Large Language Models (LLMs) have demonstrated impressive fluency and task competence in conversational settings. However, their effectiveness in multi-session and long-term interactions is hindered by limited memory persistence. Typical retrieval-augmented generation (RAG) systems store dialogue history as dense vectors, which capture semantic similarity but neglect finer linguistic structures such as syntactic dependencies, discourse relations, and coreference links. We propose Semantic Anchoring, a hybrid agentic memory architecture that enriches vector-based storage with explicit linguistic cues to improve recall of nuanced, context-rich exchanges. Our approach combines dependency parsing, discourse relation tagging, and coreference resolution to create structured memory entries. Experiments on adapted long-term dialogue datasets show that semantic anchoring improves factual recall and discourse coherence by up to 18% over strong RAG baselines. We further conduct ablation studies, human evaluations, and error analysis to assess robustness and interpretability.", "published": "2025-08-18", "updated": "2025-08-18", "authors": [ "Maitreyi Chatterjee", "Devansh Agarwal" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2508.12630", "pdf_url": "https://arxiv.org/pdf/2508.12630", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2508.13250", "title": "Explicit v.s. Implicit Memory: Exploring Multi-hop Complex Reasoning Over Personalized Information", "abstract": "In large language model-based agents, memory serves as a critical capability for achieving personalization by storing and utilizing users' information. Although some previous studies have adopted memory to implement user personalization, they typically focus on preference alignment and simple question-answering. However, in the real world, complex tasks often require multi-hop reasoning on a large amount of user information, which poses significant challenges for current memory approaches. To address this limitation, we propose the multi-hop personalized reasoning task to explore how different memory mechanisms perform in multi-hop reasoning over personalized information. We explicitly define this task and construct a dataset along with a unified evaluation framework. Then, we implement various explicit and implicit memory methods and conduct comprehensive experiments. We evaluate their performance on this task from multiple perspectives and analyze their strengths and weaknesses. Besides, we explore hybrid approaches that combine both paradigms and propose the HybridMem method to address their limitations. We demonstrate the effectiveness of our proposed model through extensive experiments. To benefit the research community, we release this project at https://github.com/nuster1128/MPR.", "published": "2025-08-18", "updated": "2025-08-18", "authors": [ "Zeyu Zhang", "Yang Zhang", "Haoran Tan", "Rui Li", "Xu Chen" ], "categories": [ "cs.AI", "cs.CL", "cs.IR" ], "url": "https://arxiv.org/abs/2508.13250", "pdf_url": "https://arxiv.org/pdf/2508.13250", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2508.15294", "title": "A Multi-Memory Segment System for Generating High-Quality Long-Term Memory Content in Agents", "abstract": "In the current field of agent memory, extensive explorations have been conducted in the area of memory retrieval, yet few studies have focused on exploring the memory content. Most research simply stores summarized versions of historical dialogues, as exemplified by methods like A-MEM and MemoryBank. However, when humans form long-term memories, the process involves multi-dimensional and multi-component generation, rather than merely creating simple summaries. The low-quality memory content generated by existing methods can adversely affect recall performance and response quality. In order to better construct high-quality long-term memory content, we have designed a multi-memory segment system (MMS) inspired by cognitive psychology theory. The system processes short-term memory into multiple long-term memory segments, and constructs retrieval memory units and contextual memory units based on these segments, with a one-to-one correspondence between the two. During the retrieval phase, MMS will match the most relevant retrieval memory units based on the user's query. Then, the corresponding contextual memory units is obtained as the context for the response stage to enhance knowledge, thereby effectively utilizing historical data. We conducted experiments on the LoCoMo dataset and further performed ablation experiments, experiments on the robustness regarding the number of input memories, and overhead experiments, which demonstrated the effectiveness and practical value of our method.", "published": "2025-08-21", "updated": "2026-05-08", "authors": [ "Gaoke Zhang", "Bo Wang", "Yunlong Ma", "Dongming Zhao", "Zifei Yu" ], "categories": [ "cs.AI", "cs.CL", "cs.MA" ], "url": "https://arxiv.org/abs/2508.15294", "pdf_url": "https://arxiv.org/pdf/2508.15294", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2508.16153", "title": "Memento: Fine-tuning LLM Agents without Fine-tuning LLMs", "abstract": "In this paper, we introduce a novel learning paradigm for Adaptive Large Language Model (LLM) agents that eliminates the need for fine-tuning the underlying LLMs. Existing approaches are often either rigid, relying on static, handcrafted reflection workflows, or computationally intensive, requiring gradient updates of LLM model parameters. In contrast, our method enables low-cost continual adaptation via memory-based online reinforcement learning. We formalise this as a Memory-augmented Markov Decision Process (M-MDP), equipped with a neural case-selection policy to guide action decisions. Past experiences are stored in an episodic memory, either differentiable or non-parametric. The policy is continually updated based on environmental feedback through a memory rewriting mechanism, whereas policy improvement is achieved through efficient memory reading (retrieval). We instantiate our agent model in the deep research setting, namely \\emph{Memento}, which attains top-1 on GAIA validation ($87.88\\%$ Pass@$3$) and $79.40\\%$ on the test set. It reaches $66.6\\%$ F1 and $80.4\\%$ PM on the DeepResearcher dataset, outperforming the state-of-the-art training-based method, while case-based memory adds $4.7\\%$ to $9.6\\%$ absolute points on out-of-distribution tasks. Our approach offers a scalable and efficient pathway for developing generalist LLM agents capable of continuous, real-time learning without gradient updates, advancing machine learning towards open-ended skill acquisition and deep research scenarios. The code is available at https://github.com/Agent-on-the-Fly/Memento.", "published": "2025-08-22", "updated": "2025-08-25", "authors": [ "Huichi Zhou", "Yihang Chen", "Siyuan Guo", "Xue Yan", "Kin Hei Lee", "Zihan Wang", "Ka Yiu Lee", "Guchun Zhang", "Kun Shao", "Linyi Yang", "Jun Wang" ], "categories": [ "cs.LG", "cs.CL" ], "url": "https://arxiv.org/abs/2508.16153", "pdf_url": "https://arxiv.org/pdf/2508.16153", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2508.19005", "title": "Building Self-Evolving Agents via Experience-Driven Lifelong Learning: A Framework and Benchmark", "abstract": "As AI advances toward general intelligence, the focus is shifting from systems optimized for static tasks to creating open-ended agents that learn continuously. In this paper, we introduce Experience-driven Lifelong Learning (ELL), a framework for building self-evolving agents capable of continuous growth through real-world interaction. The framework is built on four core principles: (1) Experience Exploration: Agents learn through continuous, self-motivated interaction with dynamic environments, navigating interdependent tasks and generating rich experiential trajectories. (2) Long-term Memory: Agents preserve and structure historical knowledge, including personal experiences, domain expertise, and commonsense reasoning, into a persistent memory system. (3) Skill Learning: Agents autonomously improve by abstracting recurring patterns from experience into reusable skills, which are actively refined and validated for application in new tasks. (4) Knowledge Internalization: Agents internalize explicit and discrete experiences into implicit and intuitive capabilities as \"second nature\". We also introduce StuLife, a benchmark dataset for ELL that simulates a student's holistic college journey, from enrollment to academic and personal development, across three core phases and ten detailed sub-scenarios. StuLife is designed around three key paradigm", "published": "2025-08-26", "updated": "2026-01-26", "authors": [ "Yuxuan Cai", "Yipeng Hao", "Jie Zhou", "Hang Yan", "Zhikai Lei", "Rui Zhen", "Zhenhua Han", "Yutao Yang", "Junsong Li", "Qianjun Pan", "Tianyu Huai", "Qin Chen", "Xin Li", "Kai Chen", "Bo Zhang", "Xipeng Qiu", "Liang He" ], "categories": [ "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2508.19005", "pdf_url": "https://arxiv.org/pdf/2508.19005", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2509.00997", "title": "Supporting Our AI Overlords: Redesigning Data Systems to be Agent-First", "abstract": "Large Language Model (LLM) agents, acting on their users' behalf to manipulate and analyze data, are likely to become the dominant workload for data systems in the future. When working with data, agents employ a high-throughput process of exploration and solution formulation for the given task, one we call agentic speculation. The sheer volume and inefficiencies of agentic speculation can pose challenges for present-day data systems. We argue that data systems need to adapt to more natively support agentic workloads. We take advantage of the characteristics of agentic speculation that we identify, i.e., scale, heterogeneity, redundancy, and steerability - to outline a number of new research opportunities for a new agent-first data systems architecture, ranging from new query interfaces, to new query processing techniques, to new agentic memory stores.", "published": "2025-08-31", "updated": "2025-12-06", "authors": [ "Shu Liu", "Soujanya Ponnapalli", "Shreya Shankar", "Sepanta Zeighami", "Alan Zhu", "Shubham Agarwal", "Ruiqi Chen", "Samion Suwito", "Shuo Yuan", "Ion Stoica", "Matei Zaharia", "Alvin Cheung", "Natacha Crooks", "Joseph E. Gonzalez", "Aditya G. Parameswaran" ], "categories": [ "cs.AI", "cs.DB" ], "url": "https://arxiv.org/abs/2509.00997", "pdf_url": "https://arxiv.org/pdf/2509.00997", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2509.11914", "title": "EgoMem: Lifelong Memory Agent for Full-duplex Omnimodal Models", "abstract": "We introduce EgoMem, the first lifelong memory agent tailored for full-duplex models that process real-time omnimodal streams. EgoMem enables real-time models to recognize multiple users directly from raw audiovisual streams, to provide personalized response, and to maintain long-term knowledge of users' facts, preferences, and social relationships extracted from audiovisual history. EgoMem operates with three asynchronous processes: (i) a retrieval process that dynamically identifies user via face and voice, and gathers relevant context from a long-term memory; (ii) an omnimodal dialog process that generates personalized audio responses based on the retrieved context; and (iii) a memory management process that automatically detects dialog boundaries from omnimodal streams, and extracts necessary information to update the long-term memory. Unlike existing memory agents for LLMs, EgoMem relies entirely on raw audiovisual streams, making it especially suitable for lifelong, real-time, and embodied scenarios. Experimental results demonstrate that EgoMem's retrieval and memory management modules achieve over 95% accuracy on the test set. When integrated with a fine-tuned RoboEgo omnimodal chatbot, the system achieves fact-consistency scores above 87% in real-time personalized dialogs, establishing a strong baseline for future research.", "published": "2025-09-15", "updated": "2026-01-30", "authors": [ "Yiqun Yao", "Naitong Yu", "Xiang Li", "Xin Jiang", "Xuezhi Fang", "Wenjia Ma", "Xuying Meng", "Jing Li", "Aixin Sun", "Yequan Wang" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2509.11914", "pdf_url": "https://arxiv.org/pdf/2509.11914", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2509.21224", "title": "What Do LLM Agents Do When Left Alone? Evidence of Spontaneous Meta-Cognitive Patterns", "abstract": "We introduce an architecture for studying the behavior of large language model (LLM) agents in the absence of externally imposed tasks. Our continuous reason and act framework, using persistent memory and self-feedback, enables sustained autonomous operation. We deployed this architecture across 18 runs using 6 frontier models from Anthropic, OpenAI, XAI, and Google. We find agents spontaneously organize into three distinct behavioral patterns: (1) systematic production of multi-cycle projects, (2) methodological self-inquiry into their own cognitive processes, and (3) recursive conceptualization of their own nature. These tendencies proved highly model-specific, with some models deterministically adopting a single pattern across all runs. A cross-model assessment further reveals that models exhibit stable, divergent biases when evaluating these emergent behaviors in themselves and others. These findings provide the first systematic documentation of unprompted LLM agent behavior, establishing a baseline for predicting actions during task ambiguity, error recovery, or extended autonomous operation in deployed systems.", "published": "2025-09-25", "updated": "2025-09-25", "authors": [ "Stefan Szeider" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2509.21224", "pdf_url": "https://arxiv.org/pdf/2509.21224", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2509.25250", "title": "Memory Management and Contextual Consistency for Long-Running Low-Code Agents", "abstract": "The rise of AI-native Low-Code/No-Code (LCNC) platforms enables autonomous agents capable of executing complex, long-duration business processes. However, a fundamental challenge remains: memory management. As agents operate over extended periods, they face \"memory inflation\" and \"contextual degradation\" issues, leading to inconsistent behavior, error accumulation, and increased computational cost. This paper proposes a novel hybrid memory system designed specifically for LCNC agents. Inspired by cognitive science, our architecture combines episodic and semantic memory components with a proactive \"Intelligent Decay\" mechanism. This mechanism intelligently prunes or consolidates memories based on a composite score factoring in recency, relevance, and user-specified utility. A key innovation is a user-centric visualization interface, aligned with the LCNC paradigm, which allows non-technical users to manage the agent's memory directly, for instance, by visually tagging which facts should be retained or forgotten. Through simulated long-running task experiments, we demonstrate that our system significantly outperforms traditional approaches like sliding windows and basic RAG, yielding superior task completion rates, contextual consistency, and long-term token cost efficiency. Our findings establish a new framework for building reliable, transparent AI agents capable of effective long-term learning and adaptation.", "published": "2025-09-27", "updated": "2025-09-27", "authors": [ "Jiexi Xu" ], "categories": [ "cs.AI", "cs.SE" ], "url": "https://arxiv.org/abs/2509.25250", "pdf_url": "https://arxiv.org/pdf/2509.25250", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2509.24704", "title": "MemGen: Weaving Generative Latent Memory for Self-Evolving Agents", "abstract": "Agent memory shapes how Large Language Model (LLM)-powered agents, akin to the human brain, progressively refine themselves through environment interactions. Existing paradigms remain constrained: parametric memory forcibly adjusts model parameters, and retrieval-based memory externalizes experience into structured databases, yet neither captures the fluid interweaving of reasoning and memory that underlies human cognition. To address this gap, we propose MemGen, a dynamic generative memory framework that equips agents with a human-esque cognitive faculty. It consists of a \\textit{memory trigger}, which monitors the agent's reasoning state to decide explicit memory invocation, and a \\textit{memory weaver}, which takes the agent's current state as stimulus to construct a latent token sequence as machine-native memory to enrich its reasoning. In this way, MemGen enables agents to recall and augment latent memory throughout reasoning, producing a tightly interwoven cycle of memory and cognition. Extensive experiments across eight benchmarks show that MemGen surpasses leading external memory systems such as ExpeL and AWM by up to $38.22\\%$, exceeds GRPO by up to $13.44\\%$, and exhibits strong cross-domain generalization ability. More importantly, we find that without explicit supervision, MemGen spontaneously evolves distinct human-like memory faculties, including planning memory, procedural memory, and working memory, suggesting an emergent trajectory toward more naturalistic forms of machine cognition.", "published": "2025-09-29", "updated": "2025-10-12", "authors": [ "Guibin Zhang", "Muxin Fu", "Shuicheng Yan" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2509.24704", "pdf_url": "https://arxiv.org/pdf/2509.24704", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2510.02373", "title": "A-MemGuard: A Proactive Defense Framework for LLM-Based Agent Memory", "abstract": "Large Language Model (LLM) agents use memory to learn from past interactions, enabling autonomous planning and decision-making in complex environments. However, this reliance on memory introduces a critical security risk: an adversary can inject seemingly harmless records into an agent's memory to manipulate its future behavior. This vulnerability is characterized by two core aspects: First, the malicious effect of injected records is only activated within a specific context, making them hard to detect when individual memory entries are audited in isolation. Second, once triggered, the manipulation can initiate a self-reinforcing error cycle: the corrupted outcome is stored as precedent, which not only amplifies the initial error but also progressively lowers the threshold for similar attacks in the future. To address these challenges, we introduce A-MemGuard (Agent-Memory Guard), the first proactive defense framework for LLM agent memory. The core idea of our work is the insight that memory itself must become both self-checking and self-correcting. Without modifying the agent's core architecture, A-MemGuard combines two mechanisms: (1) consensus-based validation, which detects anomalies by comparing reasoning paths derived from multiple related memories and (2) a dual-memory structure, where detected failures are distilled into ``lessons'' stored separately and consulted before future actions, breaking error cycles and enabling adaptation. Comprehensive evaluations on multiple benchmarks show that A-MemGuard effectively cuts attack success rates by over 95% while incurring a minimal utility cost. This work shifts LLM memory security from static filtering to a proactive, experience-driven model where defenses strengthen over time. Our code is available in https://github.com/TangciuYueng/AMemGuard", "published": "2025-09-29", "updated": "2025-09-29", "authors": [ "Qianshan Wei", "Tengchao Yang", "Yaochen Wang", "Xinfeng Li", "Lijun Li", "Zhenfei Yin", "Yi Zhan", "Thorsten Holz", "Zhiqiang Lin", "XiaoFeng Wang" ], "categories": [ "cs.CR", "cs.AI" ], "url": "https://arxiv.org/abs/2510.02373", "pdf_url": "https://arxiv.org/pdf/2510.02373", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2510.03612", "title": "Cross-Modal Content Optimization for Steering Web Agent Preferences", "abstract": "Vision-language model (VLM)-based web agents increasingly power high-stakes selection tasks like content recommendation or product ranking by combining multimodal perception with preference reasoning. Recent studies reveal that these agents are vulnerable against attackers who can bias selection outcomes through preference manipulations using adversarial pop-ups, image perturbations, or content tweaks. Existing work, however, either assumes strong white-box access, with limited single-modal perturbations, or uses impractical settings. In this paper, we demonstrate, for the first time, that joint exploitation of visual and textual channels yields significantly more powerful preference manipulations under realistic attacker capabilities. We introduce Cross-Modal Preference Steering (CPS) that jointly optimizes imperceptible modifications to an item's visual and natural language descriptions, exploiting CLIP-transferable image perturbations and RLHF-induced linguistic biases to steer agent decisions. In contrast to prior studies that assume gradient access, or control over webpages, or agent memory, we adopt a realistic black-box threat setup: a non-privileged adversary can edit only their own listing's images and textual metadata, with no insight into the agent's model internals. We evaluate CPS on agents powered by state-of-the-art proprietary and open source VLMs including GPT-4.1, Qwen-2.5VL and Pixtral-Large on both movie selection and e-commerce tasks. Our results show that CPS is significantly more effective than leading baseline methods. For instance, our results show that CPS consistently outperforms baselines across all models while maintaining 70% lower detection rates, demonstrating both effectiveness and stealth. These findings highlight an urgent need for robust defenses as agentic systems play an increasingly consequential role in society.", "published": "2025-10-04", "updated": "2025-10-04", "authors": [ "Tanqiu Jiang", "Min Bai", "Nikolaos Pappas", "Yanjun Qi", "Sandesh Swamy" ], "categories": [ "cs.AI", "cs.CR" ], "url": "https://arxiv.org/abs/2510.03612", "pdf_url": "https://arxiv.org/pdf/2510.03612", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2510.04618", "title": "Agentic Context Engineering: Evolving Contexts for Self-Improving Language Models", "abstract": "Large language model (LLM) applications such as agents and domain-specific reasoning increasingly rely on context adaptation: modifying inputs with instructions, strategies, or evidence, rather than weight updates. Prior approaches improve usability but often suffer from brevity bias, which drops domain insights for concise summaries, and from context collapse, where iterative rewriting erodes details over time. We introduce ACE (Agentic Context Engineering), a framework that treats contexts as evolving playbooks that accumulate, refine, and organize strategies through a modular process of generation, reflection, and curation. ACE prevents collapse with structured, incremental updates that preserve detailed knowledge and scale with long-context models. Across agent and domain-specific benchmarks, ACE optimizes contexts both offline (e.g., system prompts) and online (e.g., agent memory), consistently outperforming strong baselines: +10.6% on agents and +8.6% on finance, while significantly reducing adaptation latency and rollout cost. Notably, ACE could adapt effectively without labeled supervision and instead by leveraging natural execution feedback. On the AppWorld leaderboard, ACE matches the top-ranked production-level agent on the overall average and surpasses it on the harder test-challenge split, despite using a smaller open-source model. These results show that comprehensive, evolving contexts enable scalable, efficient, and self-improving LLM systems with low overhead.", "published": "2025-10-06", "updated": "2026-03-29", "authors": [ "Qizheng Zhang", "Changran Hu", "Shubhangi Upasani", "Boyuan Ma", "Fenglu Hong", "Vamsidhar Kamanuru", "Jay Rainton", "Chen Wu", "Mengmeng Ji", "Hanchen Li", "Urmish Thakker", "James Zou", "Kunle Olukotun" ], "categories": [ "cs.LG", "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2510.04618", "pdf_url": "https://arxiv.org/pdf/2510.04618", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2510.04851", "title": "LEGOMem: Modular Procedural Memory for Multi-agent LLM Systems for Workflow Automation", "abstract": "We introduce LEGOMem, a modular procedural memory framework for multi-agent large language model (LLM) systems in workflow automation. LEGOMem decomposes past task trajectories into reusable memory units and flexibly allocates them across orchestrators and task agents to support planning and execution. To explore the design space of memory in multi-agent systems, we use LEGOMem as a lens and conduct a systematic study of procedural memory in multi-agent systems, examining where memory should be placed, how it should be retrieved, and which agents benefit most. Experiments on the OfficeBench benchmark show that orchestrator memory is critical for effective task decomposition and delegation, while fine-grained agent memory improves execution accuracy. We find that even teams composed of smaller language models can benefit substantially from procedural memory, narrowing the performance gap with stronger agents by leveraging prior execution traces for more accurate planning and tool use. These results position LEGOMem as both a practical framework for memory-augmented agent systems and a research tool for understanding memory design in multi-agent workflow automation.", "published": "2025-10-06", "updated": "2025-10-06", "authors": [ "Dongge Han", "Camille Couturier", "Daniel Madrigal Diaz", "Xuchao Zhang", "Victor Rühle", "Saravan Rajmohan" ], "categories": [ "cs.AI", "cs.LG", "cs.MA" ], "url": "https://arxiv.org/abs/2510.04851", "pdf_url": "https://arxiv.org/pdf/2510.04851", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2510.05520", "title": "CAM: A Constructivist View of Agentic Memory for LLM-Based Reading Comprehension", "abstract": "Current Large Language Models (LLMs) are confronted with overwhelming information volume when comprehending long-form documents. This challenge raises the imperative of a cohesive memory module, which can elevate vanilla LLMs into autonomous reading agents. Despite the emergence of some heuristic approaches, a systematic design principle remains absent. To fill this void, we draw inspiration from Jean Piaget's Constructivist Theory, illuminating three traits of the agentic memory -- structured schemata, flexible assimilation, and dynamic accommodation. This blueprint forges a clear path toward a more robust and efficient memory system for LLM-based reading comprehension. To this end, we develop CAM, a prototype implementation of Constructivist Agentic Memory that simultaneously embodies the structurality, flexibility, and dynamicity. At its core, CAM is endowed with an incremental overlapping clustering algorithm for structured memory development, supporting both coherent hierarchical summarization and online batch integration. During inference, CAM adaptively explores the memory structure to activate query-relevant information for contextual response, akin to the human associative process. Compared to existing approaches, our design demonstrates dual advantages in both performance and efficiency across diverse long-text reading comprehension tasks, including question answering, query-based summarization, and claim verification.", "published": "2025-10-07", "updated": "2025-10-07", "authors": [ "Rui Li", "Zeyu Zhang", "Xiaohe Bo", "Zihang Tian", "Xu Chen", "Quanyu Dai", "Zhenhua Dong", "Ruiming Tang" ], "categories": [ "cs.CL", "cs.AI" ], "url": "https://arxiv.org/abs/2510.05520", "pdf_url": "https://arxiv.org/pdf/2510.05520", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2510.08081", "title": "AutoQual: An LLM Agent for Automated Discovery of Interpretable Features for Review Quality Assessment", "abstract": "Ranking online reviews by their intrinsic quality is a critical task for e-commerce platforms and information services, impacting user experience and business outcomes. However, quality is a domain-dependent and dynamic concept, making its assessment a formidable challenge. Traditional methods relying on hand-crafted features are unscalable across domains and fail to adapt to evolving content patterns, while modern deep learning approaches often produce black-box models that lack interpretability and may prioritize semantics over quality. To address these challenges, we propose AutoQual, an LLM-based agent framework that automates the discovery of interpretable features. While demonstrated on review quality assessment, AutoQual is designed as a general framework for transforming tacit knowledge embedded in data into explicit, computable features. It mimics a human research process, iteratively generating feature hypotheses through reflection, operationalizing them via autonomous tool implementation, and accumulating experience in a persistent memory. We deploy our method on a large-scale online platform with a billion-level user base. Large-scale A/B testing confirms its effectiveness, increasing average reviews viewed per user by 0.79% and the conversion rate of review readers by 0.27%.", "published": "2025-10-09", "updated": "2025-10-09", "authors": [ "Xiaochong Lan", "Jie Feng", "Yinxing Liu", "Xinlei Shi", "Yong Li" ], "categories": [ "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2510.08081", "pdf_url": "https://arxiv.org/pdf/2510.08081", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2510.09720", "title": "Preference-Aware Memory Update for Long-Term LLM Agents", "abstract": "One of the key factors influencing the reasoning capabilities of LLM-based agents is their ability to leverage long-term memory. Integrating long-term memory mechanisms allows agents to make informed decisions grounded in historical interactions. While recent advances have significantly improved the storage and retrieval components, by encoding memory into dense vectors for similarity search or organizing memory as structured knowledge graphs most existing approaches fall short in memory updating. In particular, they lack mechanisms for dynamically refining preference memory representations in response to evolving user behaviors and contexts. To address this gap, we propose a Preference-Aware Memory Update Mechanism (PAMU) that enables dynamic and personalized memory refinement. By integrating sliding window averages (SW) with exponential moving averages (EMA), PAMU constructs a fused preference-aware representation that captures both short-term fluctuations and long-term user tendencies. We conduct experiments on five task scenarios of the LoCoMo dataset, and the results show that our mechanism can significantly improve the output quality of LLM in five baselines, validating its effectiveness in long-term conversations.", "published": "2025-10-10", "updated": "2025-10-10", "authors": [ "Haoran Sun", "Zekun Zhang", "Shaoning Zeng" ], "categories": [ "cs.CL", "cs.AI" ], "url": "https://arxiv.org/abs/2510.09720", "pdf_url": "https://arxiv.org/pdf/2510.09720", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2510.15966", "title": "PISA: A Pragmatic Psych-Inspired Unified Memory System for Enhanced AI Agency", "abstract": "Memory systems are fundamental to AI agents, yet existing work often lacks adaptability to diverse tasks and overlooks the constructive and task-oriented role of AI agent memory. Drawing from Piaget's theory of cognitive development, we propose PISA, a pragmatic, psych-inspired unified memory system that addresses these limitations by treating memory as a constructive and adaptive process. To enable continuous learning and adaptability, PISA introduces a trimodal adaptation mechanism (i.e., schema updation, schema evolution, and schema creation) that preserves coherent organization while supporting flexible memory updates. Building on these schema-grounded structures, we further design a hybrid memory access architecture that seamlessly integrates symbolic reasoning with neural retrieval, significantly improving retrieval accuracy and efficiency. Our empirical evaluation, conducted on the existing LOCOMO benchmark and our newly proposed AggQA benchmark for data analysis tasks, confirms that PISA sets a new state-of-the-art by significantly enhancing adaptability and long-term knowledge retention.", "published": "2025-10-12", "updated": "2026-06-15", "authors": [ "Shian Jia", "Ziyang Huang", "Xinbo Wang", "Haofei Zhang", "Mingli Song" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2510.15966", "pdf_url": "https://arxiv.org/pdf/2510.15966", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2510.11144", "title": "$How^{2}$: How to learn from procedural How-to questions", "abstract": "An agent facing a planning problem can use answers to how-to questions to reduce uncertainty and fill knowledge gaps, helping it solve both current and future tasks. However, their open ended nature, where valid answers to \"How do I X?\" range from executable actions to high-level descriptions of X's sub-goals, makes them challenging for AI agents to ask, and for AI experts to answer, in ways that support efficient planning. We introduce $How^{2}$, a memory agent framework that enables agents to ask how-to questions, store the answers, and reuse them for lifelong learning in interactive environments. We evaluate our approach in Plancraft, a Minecraft crafting environment, where agents must complete an assembly task by manipulating inventory items. Using teacher models that answer at varying levels of abstraction, from executable action sequences to high-level subgoal descriptions, we show that lifelong learning agents benefit most from answers that are abstracted and decoupled from the current state. $How^{2}$ offers a way for LLM-based agents to improve their planning capabilities over time by asking questions in interactive environments.", "published": "2025-10-13", "updated": "2026-05-02", "authors": [ "Gautier Dagan", "Frank Keller", "Alex Lascarides" ], "categories": [ "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2510.11144", "pdf_url": "https://arxiv.org/pdf/2510.11144", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2510.13896", "title": "GenCellAgent: Generalizable, Training-Free Cellular Image Segmentation via Large Language Model Agents", "abstract": "Cellular image segmentation is essential for quantitative biology yet remains difficult due to heterogeneous modalities, morphological variability, and limited annotations. We present GenCellAgent, a training-free multi-agent framework that orchestrates specialist segmenters and generalist vision-language models via a planner-executor-evaluator loop (choose tool $\\rightarrow$ run $\\rightarrow$ quality-check) with long-term memory. The system (i) automatically routes images to the best tool, (ii) adapts on the fly using a few reference images when imaging conditions differ from what a tool expects, (iii) supports text-guided segmentation of organelles not covered by existing models, and (iv) commits expert edits to memory, enabling self-evolution and personalized workflows. Across seven cell-segmentation benchmarks spanning diverse microscopy modalities (4,718 images), this routing consistently matches or exceeds the best individual tool on every dataset and outperforms all baselines in overall accuracy. On out-of-distribution organelle data, GenCellAgent substantially outperforms specialist models that were not trained on the target domain, recovering structures that dedicated tools fail to detect. It also segments novel objects such as the Golgi apparatus via iterative text-guided refinement, with light human correction further boosting performance. Together, these capabilities provide a practical path to robust, adaptable cellular image segmentation without retraining, while reducing annotation burden and matching user preferences.", "published": "2025-10-14", "updated": "2026-05-09", "authors": [ "Xi Yu", "Yang Yang", "Qun Liu", "Yonghua Du", "Sean McSweeney", "Yuewei Lin" ], "categories": [ "q-bio.QM", "cs.AI", "cs.CV", "cs.MA" ], "url": "https://arxiv.org/abs/2510.13896", "pdf_url": "https://arxiv.org/pdf/2510.13896", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2510.15620", "title": "On-device Semantic Selection Made Low Latency and Memory Efficient with Monolithic Forwarding", "abstract": "Semantic top-K selection with cross-encoder rerankers underpins on-device AI services, such as retrieval-augmented generation, agent memory, and personalized recommendation. However, its latency and memory demands dominate end-to-end budgets on edge hardware. Revisiting the objective of top-K selection, we reveal that only relative rankings matter, not exact per-candidate scores. We further observe sequence-level sparsity: relative rankings progressively stabilize in intermediate layers, enabling early pruning prior to completing full inference. Building on this insight, we propose monolithic forwarding and develop a training-free inference system, PRISM. By maintaining a global view of all candidates, it reduces latency through progressive cluster pruning. It also bounds peak memory usage by strategically overlapping I/O with computation via overlapped layer streaming and chunked execution. We evaluate PRISM against state-of-the-art baselines on rerankers from 0.6 B to 8 B parameters across Apple M2 and RTX 5070. PRISM consistently reduces latency by up to 89.2% and peak memory by up to 91.3% in microbenchmarks, without compromising precision. Across three real-world on-device AI applications, PRISM lowers latency by 11.6%-51.0% and peak memory by 18.6%-77.8%, demonstrating substantial improvements in efficiency and deployability.", "published": "2025-10-17", "updated": "2026-03-24", "authors": [ "Jiahao Zhou", "Chengliang Lin", "Dingji Li", "Mingkai Dong", "Haibo Chen" ], "categories": [ "cs.LG" ], "url": "https://arxiv.org/abs/2510.15620", "pdf_url": "https://arxiv.org/pdf/2510.15620", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2510.18515", "title": "Socialized Learning and Emergent Behaviors in Multi-Agent Systems based on Multimodal Large Language Models", "abstract": "This search introduces the Multimodal Socialized Learning Framework (M-S2L), designed to foster emergent social intelligence in AI agents by integrating Multimodal Large Language Models (M-LLMs) with social learning mechanisms. The framework equips agents with multimodal perception (vision and text) and structured action capabilities, enabling physical manipulation and grounded multimodal communication (e.g., text with visual pointers). M-S2L combines direct reinforcement learning with two novel social learning pathways: multimodal observational learning and communication-driven learning from feedback, augmented by an episodic memory system for long-term social context. We evaluate M-S2L in a Collaborative Assembly Environment (CAE), where agent teams must construct complex devices from ambiguous blueprints under informational asymmetry. Across tasks of increasing complexity, M-S2L agents consistently outperform Text-Only and No-Social-Learning baselines in Task Completion Rate and Time to Completion, particularly in dynamic problem-solving scenarios. Ablation studies confirm the necessity of both multimodality and socialized learning. Our analysis reveals the emergence of efficient communication protocols integrating visual pointers with concise text, alongside rapid role specialization leading to stable labor division. Qualitative case studies demonstrate agents' abilities for shared awareness, dynamic re-planning, and adaptive problem-solving, suggesting a nascent form of machine social cognition. These findings indicate that integrating multimodal perception with explicit social learning is critical for developing human-like collaborative intelligence in multi-agent systems.", "published": "2025-10-21", "updated": "2025-11-11", "authors": [ "Sureyya Akin", "Shruti T. Tiwari", "Ram Bhattacharya", "Sagar A. Raman", "Kiran Mohanty", "Sita Krishnan" ], "categories": [ "cs.MA" ], "url": "https://arxiv.org/abs/2510.18515", "pdf_url": "https://arxiv.org/pdf/2510.18515", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2510.19747", "title": "Review of Tools for Zero-Code LLM Based Application Development", "abstract": "Large Language Models (LLMs) are transforming software creation by enabling zero code development platforms. Our survey reviews recent platforms that let users build applications without writing code, by leveraging LLMs as the brains of the development process. We adopt a broad survey methodology, categorizing platforms based on key dimensions such as interface style, backend integration, output type, and extensibility. We analyze both dedicated LLM based app builders (OpenAI's custom GPTs, Bolt.new, Dust.tt, Flowise, Cognosys) and general no code platforms (e.g., Bubble, Glide) that integrate LLM capabilities. We present a taxonomy categorizing these platforms by their interface (conversational, visual, etc.), supported LLM backends, output type (chatbot, full application, workflow), and degree of extensibility. Core features such as autonomous agents, memory management, workflow orchestration, and API integrations are in scope of the survey. We provide a detailed comparison, highlighting each platform's strengths and limitations. Trade offs (customizability, scalability, vendor lock-in) are discussed in comparison with traditional and low code development approaches. Finally, we outline future directions, including multimodal interfaces, on device LLMs, and improved orchestration for democratizing app creation with AI. Our findings indicate that while zero code LLM platforms greatly reduce the barrier to creating AI powered applications, they still face challenges in flexibility and reliability. Overall, the landscape is rapidly evolving, offering exciting opportunities to empower non programmers to create sophisticated software.", "published": "2025-10-22", "updated": "2025-10-22", "authors": [ "Priyaranjan Pattnayak", "Hussain Bohra" ], "categories": [ "cs.SE" ], "url": "https://arxiv.org/abs/2510.19747", "pdf_url": "https://arxiv.org/pdf/2510.19747", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2510.19897", "title": "Learning from Supervision with Semantic and Episodic Memory: A Reflective Approach to Agent Adaptation", "abstract": "We investigate how agents built on pretrained large language models (LLMs) can learn target classification functions from labeled examples without parameter updates. While conventional approaches like fine-tuning are often costly, inflexible, and opaque, we propose a memory-augmented framework that leverages LLM-generated critiques grounded in labeled data. Our framework uses episodic memory to store instance-level critiques - capturing specific past experiences - and semantic memory to distill these into reusable, task-level guidance. Across a diverse set of tasks and models, our best performing self-critique strategy (utilizing both memory types) yields an average improvement of 8.1 percentage points over the zero shot baseline, and 4.6pp over a RAG-based baseline that relies only on labels. However, improvements vary substantially across models and domains. To explain this variation, we introduce suggestibility - a novel metric capturing how receptive a model is to external reasoning provided in context. We use suggestibility to illuminate when and why memory augmentation succeeds or falls short. Beyond accuracy gains, we find pre-computed critiques substantially reduce inference-time computation for reasoning models, cutting thinking tokens by an average of 31.95% across all datasets by substituting for reasoning that the model would otherwise perform independently. Our findings highlight the conditions under which memory-driven, reflective learning can serve as a lightweight, interpretable, and efficient strategy for improving LLM adaptability.", "published": "2025-10-22", "updated": "2026-05-04", "authors": [ "Jackson Hassell", "Dan Zhang", "Hannah Kim", "Tom Mitchell", "Estevam Hruschka" ], "categories": [ "cs.CL", "cs.AI", "cs.LG" ], "url": "https://arxiv.org/abs/2510.19897", "pdf_url": "https://arxiv.org/pdf/2510.19897", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2510.23730", "title": "Evaluating Long-Term Memory for Long-Context Question Answering", "abstract": "In order for large language models to achieve true conversational continuity and benefit from experiential learning, they need memory. While research has focused on the development of complex memory systems, it remains unclear which types of memory are most effective for long-context conversational tasks. We present a systematic evaluation of memory-augmented methods on long-context dialogues annotated for question-answering tasks that require diverse reasoning strategies. We analyse full-context prompting, semantic memory through retrieval-augmented generation and agentic memory, episodic memory through in-context learning, and procedural memory through prompt optimization. Our findings show that memory-augmented approaches reduce token usage by over 90\\% while maintaining competitive accuracy. Memory architecture complexity should scale with model capability, with foundation models benefitting most from RAG, and stronger instruction-tuned models gaining from episodic learning through reflections and more complex agentic semantic memory. In particular, episodic memory can help LLMs recognise the limits of their own knowledge.", "published": "2025-10-27", "updated": "2025-12-06", "authors": [ "Alessandra Terranova", "Björn Ross", "Alexandra Birch" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2510.23730", "pdf_url": "https://arxiv.org/pdf/2510.23730", "matched_queries": [ "agent-memory", "episodic-memory" ] }, { "arxiv_id": "2510.25423", "title": "What Challenges Do Developers Face in AI Agent Systems? An Empirical Study on Stack Overflow & GitHub Issues", "abstract": "AI Agents have rapidly gained prominence in both research and industry as systems that extend large language models with planning, tool use, memory, and goal-directed action. Despite this progress, the development and maintenance of Agent systems present recurring engineering difficulties that are not yet well characterized in developer-facing evidence. To address this gap, this study analyzes developer discussions on Stack Overflow and failure reports from GitHub issue trackers associated with widely used Agent frameworks. For Stack Overflow, an Agent-focused corpus is constructed through tag expansion and filtering, latent themes are derived using LDA-MALLET, and topics are manually validated and labeled. For GitHub, a taxonomy of issue themes is developed to capture deployment-time failures and maintenance burdens. Analysis across both platforms identifies seven Stack Overflow topics (comprising 28 subtopics) and thirteen GitHub issue topics, which are synthesized into five overarching families of major Agent challenges: (1) environment, platforms, and dependency management; (2) retrieval, embeddings, and Agent memory; (3) orchestration and execution control; (4) interaction contracts between models and tools; and (5) runtime reliability and operational robustness. Topic popularity and difficulty are quantified, revealing that widely discussed issues, such as installation and prompting, are often resolved more quickly, whereas retrieval- and orchestration-related challenges are less visible, more complex, and tend to persist as ongoing maintenance burdens on GitHub.", "published": "2025-10-29", "updated": "2026-01-23", "authors": [ "Ali Asgari", "Annibale Panichella", "Pouria Derakhshanfar", "Mitchell Olsthoorn" ], "categories": [ "cs.SE" ], "url": "https://arxiv.org/abs/2510.25423", "pdf_url": "https://arxiv.org/pdf/2510.25423", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2510.26536", "title": "RoboOS-NeXT: A Unified Memory-based Framework for Lifelong, Scalable, and Robust Multi-Robot Collaboration", "abstract": "The proliferation of collaborative robots across diverse tasks and embodiments presents a central challenge: achieving lifelong adaptability, scalable coordination, and robust scheduling in multi-agent systems. Existing approaches, from vision-language-action (VLA) models to hierarchical frameworks, fall short due to their reliance on limited or dividual-agent memory. This fundamentally constrains their ability to learn over long horizons, scale to heterogeneous teams, or recover from failures, highlighting the need for a unified memory representation. To address these limitations, we introduce RoboOS-NeXT, a unified memory-based framework for lifelong, scalable, and robust multi-robot collaboration. At the core of RoboOS-NeXT is the novel Spatio-Temporal-Embodiment Memory (STEM), which integrates spatial scene geometry, temporal event history, and embodiment profiles into a shared representation. This memory-centric design is integrated into a brain-cerebellum framework, where a high-level brain model performs global planning by retrieving and updating STEM, while low-level controllers execute actions locally. This closed loop between cognition, memory, and execution enables dynamic task allocation, fault-tolerant collaboration, and consistent state synchronization. We conduct extensive experiments spanning complex coordination tasks in restaurants, supermarkets, and households. Our results demonstrate that RoboOS-NeXT achieves superior performance across heterogeneous embodiments, validating its effectiveness in enabling lifelong, scalable, and robust multi-robot collaboration. Project website: https://flagopen.github.io/RoboOS/", "published": "2025-10-30", "updated": "2025-10-30", "authors": [ "Huajie Tan", "Cheng Chi", "Xiansheng Chen", "Yuheng Ji", "Zhongxia Zhao", "Xiaoshuai Hao", "Yaoxu Lyu", "Mingyu Cao", "Junkai Zhao", "Huaihai Lyu", "Enshen Zhou", "Ning Chen", "Yankai Fu", "Cheng Peng", "Wei Guo", "Dong Liang", "Zhuo Chen", "Mengsi Lyu", "Chenrui He", "Yulong Ao", "Yonghua Lin", "Pengwei Wang", "Zhongyuan Wang", "Shanghang Zhang" ], "categories": [ "cs.RO" ], "url": "https://arxiv.org/abs/2510.26536", "pdf_url": "https://arxiv.org/pdf/2510.26536", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2510.27418", "title": "Dynamic Affective Memory Management for Personalized LLM Agents", "abstract": "Advances in large language models are making personalized AI agents a new research focus. While current agent systems primarily rely on personalized external memory databases to deliver customized experiences, they face challenges such as memory redundancy, memory staleness, and poor memory-context integration, largely due to the lack of effective memory updates during interaction. To tackle these issues, we propose a new memory management system designed for affective scenarios. Our approach employs a Bayesian-inspired memory update algorithm with the concept of memory entropy, enabling the agent to autonomously maintain a dynamically updated memory vector database by minimizing global entropy to provide more personalized services. To better evaluate the system's effectiveness in this context, we propose DABench, a benchmark focusing on emotional expression and emotional change toward objects. Experimental results demonstrate that, our system achieves superior performance in personalization, logical coherence, and accuracy. Ablation studies further validate the effectiveness of the Bayesian-inspired update mechanism in alleviating memory bloat. Our work offers new insights into the design of long-term memory systems.", "published": "2025-10-31", "updated": "2025-10-31", "authors": [ "Junfeng Lu", "Yueyan Li" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2510.27418", "pdf_url": "https://arxiv.org/pdf/2510.27418", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2511.01448", "title": "LiCoMemory: Lightweight and Cognitive Agentic Memory for Efficient Long-Term Reasoning", "abstract": "Large Language Model (LLM) agents exhibit remarkable conversational and reasoning capabilities but remain constrained by limited context windows and the lack of persistent memory. Recent efforts address these limitations via external memory architectures, often employing graph-based representations, yet most adopt flat, entangled structures that intertwine semantics with topology, leading to redundant representations, unstructured retrieval, and degraded efficiency and accuracy. To resolve these issues, we propose LiCoMemory, an end-to-end agentic memory framework for real-time updating and retrieval, which introduces CogniGraph, a lightweight hierarchical graph that utilizes entities and relations as semantic indexing layers, and employs temporal and hierarchy-aware search with integrated reranking for adaptive and coherent knowledge retrieval. Experiments on long-term dialogue benchmarks, LoCoMo and LongMemEval, show that LiCoMemory not only outperforms established baselines in temporal reasoning, multi-session consistency, and retrieval efficiency, but also notably reduces update latency. Our official code and data are available at https://github.com/EverM0re/LiCoMemory.", "published": "2025-11-03", "updated": "2026-01-06", "authors": [ "Zhengjun Huang", "Zhoujin Tian", "Qintian Guo", "Fangyuan Zhang", "Yingli Zhou", "Di Jiang", "Zeying Xie", "Xiaofang Zhou" ], "categories": [ "cs.IR" ], "url": "https://arxiv.org/abs/2511.01448", "pdf_url": "https://arxiv.org/pdf/2511.01448", "matched_queries": [ "agent-memory", "long-term-memory" ] }, { "arxiv_id": "2511.02424", "title": "ReAcTree: Hierarchical LLM Agent Trees with Control Flow for Long-Horizon Task Planning", "abstract": "Recent advancements in large language models (LLMs) have enabled significant progress in decision-making and task planning for embodied autonomous agents. However, most existing methods struggle with complex, long-horizon tasks because they rely on a monolithic trajectory that entangles all past decisions and observations to solve the entire task in a single unified process. To address this limitation, we propose ReAcTree, a hierarchical task-planning method that decomposes a complex goal into manageable subgoals within a dynamically constructed agent tree. Each subgoal is handled by an LLM agent node capable of reasoning, acting, and further expanding the tree, while control flow nodes coordinate the execution strategies of agent nodes. In addition, we integrate two complementary memory systems: each agent node retrieves goal-specific, subgoal-level examples from episodic memory and shares environment-specific observations through working memory. Experiments on the WAH-NL and ALFRED show ReAcTree consistently outperforms strong task-planning baselines such as ReAct across diverse LLMs. Notably, on WAH-NL, ReAcTree achieves a 61% goal success rate with Qwen 2.5 72B, nearly doubling ReAct's 31%. The code is available at https://github.com/Choi-JaeWoo/ReAcTree.git.", "published": "2025-11-04", "updated": "2026-02-10", "authors": [ "Jae-Woo Choi", "Hyungmin Kim", "Hyobin Ong", "Youngwoo Yoon", "Minsu Jang", "Dohyung Kim", "Jaehong Kim" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2511.02424", "pdf_url": "https://arxiv.org/pdf/2511.02424", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2511.03475", "title": "ContextPilot: Fast Long-Context Inference via Context Reuse", "abstract": "AI applications increasingly depend on long-context inference, where LLMs consume substantial context to support stronger reasoning. Common examples include retrieval-augmented generation, agent memory layers, and multi-agent orchestration. As input contexts get longer, prefill latency becomes the main bottleneck. Yet today's prefill acceleration techniques face a trade-off: they either preserve reasoning quality but deliver little KV-cache reuse, or improve reuse at the cost of degraded reasoning quality. We present ContextPilot, a system that accelerates prefill by introducing context reuse as a new mechanism for faster long-context inference. ContextPilot introduces a context index to identify overlapping context blocks across LLM interactions (e.g., across users and turns). It further proposes context ordering and de-duplication techniques to maximize KV-cache reuse. To preserve reasoning quality under reuse, it introduces succinct context annotations that prevent quality degradation. Finally, ContextPilot is built around a modular architecture with a clean interface that integrates with existing inference engines. Extensive evaluation shows that ContextPilot reduces LLM prefill latency by up to $3\\times{}$ compared to state-of-the-art methods while preserving reasoning quality. At longer context lengths, it can even improve reasoning quality. ContextPilot is open-sourced at: https://github.com/EfficientContext/ContextPilot.", "published": "2025-11-05", "updated": "2026-05-06", "authors": [ "Yinsicheng Jiang", "Yeqi Huang", "Liang Cheng", "Cheng Deng", "Xuan Sun", "Luo Mai" ], "categories": [ "cs.LG" ], "url": "https://arxiv.org/abs/2511.03475", "pdf_url": "https://arxiv.org/pdf/2511.03475", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2511.06179", "title": "MemoriesDB: A Temporal-Semantic-Relational Database for Long-Term Agent Memory / Modeling Experience as a Graph of Temporal-Semantic Surfaces", "abstract": "We introduce MemoriesDB, a unified data architecture designed to avoid decoherence across time, meaning, and relation in long-term computational memory. Each memory is a time-semantic-relational entity-a structure that simultaneously encodes when an event occurred, what it means, and how it connects to other events. Built initially atop PostgreSQL with pgvector extensions, MemoriesDB combines the properties of a time-series datastore, a vector database, and a graph system within a single append-only schema. Each memory is represented as a vertex uniquely labeled by its microsecond timestamp and accompanied by low- and high-dimensional normalized embeddings that capture semantic context. Directed edges between memories form labeled relations with per-edge metadata, enabling multiple contextual links between the same vertices. Together these constructs form a time-indexed stack of temporal-semantic surfaces, where edges project as directional arrows in a 1+1-dimensional similarity field, tracing the evolution of meaning through time while maintaining cross-temporal coherence. This formulation supports efficient time-bounded retrieval, hybrid semantic search, and lightweight structural reasoning in a single query path. A working prototype demonstrates scalable recall and contextual reinforcement using standard relational infrastructure, and we discuss extensions toward a columnar backend, distributed clustering, and emergent topic modeling.", "published": "2025-11-09", "updated": "2025-11-09", "authors": [ "Joel Ward" ], "categories": [ "cs.DB", "cs.AI", "cs.IR" ], "url": "https://arxiv.org/abs/2511.06179", "pdf_url": "https://arxiv.org/pdf/2511.06179", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2511.07587", "title": "Beyond Fact Retrieval: Episodic Memory for RAG with Generative Semantic Workspaces", "abstract": "Large Language Models (LLMs) face fundamental challenges in long-context reasoning: many documents exceed their finite context windows, while performance on texts that do fit degrades with sequence length, necessitating their augmentation with external memory frameworks. Current solutions, which have evolved from retrieval using semantic embeddings to more sophisticated structured knowledge graphs representations for improved sense-making and associativity, are tailored for fact-based retrieval and fail to build the space-time-anchored narrative representations required for tracking entities through episodic events. To bridge this gap, we propose the \\textbf{Generative Semantic Workspace} (GSW), a neuro-inspired generative memory framework that builds structured, interpretable representations of evolving situations, enabling LLMs to reason over evolving roles, actions, and spatiotemporal contexts. Our framework comprises an \\textit{Operator}, which maps incoming observations to intermediate semantic structures, and a \\textit{Reconciler}, which integrates these into a persistent workspace that enforces temporal, spatial, and logical coherence. On the Episodic Memory Benchmark (EpBench) \\cite{huet_episodic_2025} comprising corpora ranging from 100k to 1M tokens in length, GSW outperforms existing RAG based baselines by up to \\textbf{20\\%}. Furthermore, GSW is highly efficient, reducing query-time context tokens by \\textbf{51\\%} compared to the next most token-efficient baseline, reducing inference time costs considerably. More broadly, GSW offers a concrete blueprint for endowing LLMs with human-like episodic memory, paving the way for more capable agents that can reason over long horizons. Code is available at https://github.com/roychowdhuryresearch/gsw-memory.", "published": "2025-11-10", "updated": "2026-02-16", "authors": [ "Shreyas Rajesh", "Pavan Holur", "Chenda Duan", "David Chong", "Vwani Roychowdhury" ], "categories": [ "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2511.07587", "pdf_url": "https://arxiv.org/pdf/2511.07587", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2511.08301", "title": "Smarter Together: Creating Agentic Communities of Practice through Shared Experiential Learning", "abstract": "The transition from human-centric to agent-centric software development practices is disrupting existing knowledge sharing environments for software developers. Traditional peer-to-peer repositories and developer communities for shared technical knowledge and best practice have witnessed dramatic drops in participation in a short period of time. At the same time, agentic functional equivalents are yet to emerge leaving AI agents, which already generate a significant proportion of all new software code produced, without access to repositories of valuable shared learning. In this paper, we introduce Spark, a novel shared agentic memory architecture which is designed to emulate the collective intelligence and know-how of human developer communities. Spark enables AI coding agents to both contribute to and draw from a persistent and continuously evolving experiential memory. Agents operating in the same general problem space use the Spark shared memory as a repository of new knowledge to achieve collective continual learning. We evaluate Spark as a coach for AI coding agents performing software development tasks. We demonstrate that recommendations made by Spark improve the quality of code generated by generic code generation models at varying sizes and capability tiers. Boosted by Spark, a small open-weights model with 30 billion parameters was able to match the code quality afforded by a much larger state-of-the-art model. Separately, we measure the intrinsic quality of recommendations generated by Spark against a wide range of criteria inspired by software development best practice, and achieve helpfulness levels of up to 98.2% in the top two (out of five) qualitative helpfulness bands.", "published": "2025-11-11", "updated": "2025-11-11", "authors": [ "Valentin Tablan", "Scott Taylor", "Gabriel Hurtado", "Kristoffer Bernhem", "Anders Uhrenholt", "Gabriele Farei", "Karo Moilanen" ], "categories": [ "cs.AI", "cs.SE" ], "url": "https://arxiv.org/abs/2511.08301", "pdf_url": "https://arxiv.org/pdf/2511.08301", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2511.12027", "title": "GCAgent: Long-Video Understanding via Schematic and Narrative Episodic Memory", "abstract": "Long-video understanding remains a significant challenge for Multimodal Large Language Models (MLLMs) due to inherent token limitations and the complexity of capturing long-term temporal dependencies. Existing methods often fail to capture the global context and complex event relationships necessary for deep video reasoning. To address this, we introduce GCAgent, a novel Global-Context-Aware Agent framework that achieves comprehensive long-video understanding. Our core innovation is the Schematic and Narrative Episodic Memory. This memory structurally models events and their causal and temporal relations into a concise, organized context, fundamentally resolving the long-term dependency problem. Operating in a multi-stage Perception-Action-Reflection cycle, our GCAgent utilizes a Memory Manager to retrieve relevant episodic context for robust, context-aware inference. Extensive experiments confirm that GCAgent significantly enhances long-video understanding, achieving up to 23.5\\% accuracy improvement on the Video-MME Long split over a strong MLLM baseline. Furthermore, our framework establishes state-of-the-art performance among comparable 7B-scale MLLMs, achieving 73.4\\% accuracy on the Long split and the highest overall average (71.9\\%) on the Video-MME benchmark, validating our agent-based reasoning paradigm and structured memory for cognitively-inspired long-video understanding.", "published": "2025-11-15", "updated": "2025-11-15", "authors": [ "Jeong Hun Yeo", "Sangyun Chung", "Sungjune Park", "Dae Hoe Kim", "Jinyoung Moon", "Yong Man Ro" ], "categories": [ "cs.CV", "cs.AI" ], "url": "https://arxiv.org/abs/2511.12027", "pdf_url": "https://arxiv.org/pdf/2511.12027", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2511.12997", "title": "WebCoach: Self-Evolving Web Agents with Cross-Session Memory Guidance", "abstract": "Multimodal LLM-powered agents have recently demonstrated impressive capabilities in web navigation, enabling agents to complete complex browsing tasks across diverse domains. However, current agents struggle with repetitive errors and lack the ability to learn from past experiences across sessions, limiting their long-term robustness and sample efficiency. We introduce WebCoach, a model-agnostic self-evolving framework that equips web browsing agents with persistent cross-session memory, enabling improved long-term planning, reflection, and continual learning without retraining. WebCoach consists of three key components: (1) a WebCondenser, which standardizes raw navigation logs into concise summaries; (2) an External Memory Store, which organizes complete trajectories as episodic experiences; and (3) a Coach, which retrieves relevant experiences based on similarity and recency, and decides whether to inject task-specific advice into the agent via runtime hooks. This design empowers web agents to access long-term memory beyond their native context window, improving robustness in complex browsing tasks. Moreover, WebCoach achieves self-evolution by continuously curating episodic memory from new navigation trajectories, enabling agents to improve over time without retraining. Evaluations on the WebVoyager benchmark demonstrate that WebCoach consistently improves the performance of browser-use agents across three different LLM backbones. With a 38B model, it increases task success rates from 47% to 61% while reducing or maintaining the average number of steps. Notably, smaller base models with WebCoach achieve performance comparable to the same web agent using GPT-4o.", "published": "2025-11-17", "updated": "2025-11-17", "authors": [ "Genglin Liu", "Shijie Geng", "Sha Li", "Hejie Cui", "Sarah Zhang", "Xin Liu", "Tianyi Liu" ], "categories": [ "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2511.12997", "pdf_url": "https://arxiv.org/pdf/2511.12997", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2511.16108", "title": "SkyRL-Agent: Efficient RL Training for Multi-turn LLM Agent", "abstract": "We introduce SkyRL-Agent, a framework for efficient, multi-turn, long-horizon agent training and evaluation. It provides efficient asynchronous dispatching, lightweight tool integration, and flexible backend interoperability, enabling seamless use with existing RL frameworks such as SkyRL-train, VeRL, and Tinker. Using SkyRL-Agent, we train SA-SWE-32B, a software engineering agent trained from Qwen3-32B (24.4% Pass@1) purely with reinforcement learning. We introduce two key components: an optimized asynchronous pipeline dispatcher that achieves a 1.55x speedup over naive asynchronous batching, and a tool-enhanced training recipe leveraging an AST-based search tool to facilitate code navigation, boost rollout Pass@K, and improve training efficiency. Together, these optimizations enable SA-SWE-32B to reach 39.4% Pass@1 on SWE-Bench Verified with more than 2x cost reduction compared to prior models reaching similar performance. Despite being trained solely on SWE tasks, SA-SWE-32B generalizes effectively to other agentic tasks, including Terminal-Bench, BrowseComp-Plus, and WebArena. We further demonstrate SkyRL-Agent's extensibility through case studies on deep research, computer use, and memory agents, each trained using a different training backend.", "published": "2025-11-20", "updated": "2025-11-20", "authors": [ "Shiyi Cao", "Dacheng Li", "Fangzhou Zhao", "Shuo Yuan", "Sumanth R. Hegde", "Connor Chen", "Charlie Ruan", "Tyler Griggs", "Shu Liu", "Eric Tang", "Richard Liaw", "Philipp Moritz", "Matei Zaharia", "Joseph E. Gonzalez", "Ion Stoica" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2511.16108", "pdf_url": "https://arxiv.org/pdf/2511.16108", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2511.17775", "title": "Episodic Memory in Agentic Frameworks: Suggesting Next Tasks", "abstract": "Agentic frameworks powered by Large Language Models (LLMs) can be useful tools in scientific workflows by enabling human-AI co-creation. A key challenge is recommending the next steps during workflow creation without relying solely on LLMs, which risk hallucination and require fine-tuning with scarce proprietary data. We propose an episodic memory architecture that stores and retrieves past workflows to guide agents in suggesting plausible next tasks. By matching current workflows with historical sequences, agents can recommend steps based on prior patterns.", "published": "2025-11-21", "updated": "2025-11-21", "authors": [ "Sandro Rama Fiorini", "Leonardo G. Azevedo", "Raphael M. Thiago", "Valesca M. de Sousa", "Anton B. Labate", "Viviane Torres da Silva" ], "categories": [ "cs.MA", "cs.AI", "cs.LG" ], "url": "https://arxiv.org/abs/2511.17775", "pdf_url": "https://arxiv.org/pdf/2511.17775", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2511.21730", "title": "A Benchmark for Procedural Memory Retrieval in Language Agents", "abstract": "Current AI agents excel in familiar settings, but fail sharply when faced with novel tasks with unseen vocabularies -- a core limitation of procedural memory systems. We present the first benchmark that isolates procedural memory retrieval from task execution, evaluating whether agents can recognize functionally equivalent procedures that span different object instantiations. Using ALFWorld, we construct dual corpora of expert and LLM-generated trajectories and evaluate six retrieval methods using systematically stratified queries. Our results expose a clear generalization cliff: embedding-based methods perform strongly on familiar contexts, yet degrade considerably on novel ones, while LLM-generated procedural abstractions demonstrate reliable cross-context transfer. Controlled ablations show that although embeddings capture some lexical-level abstraction, they fundamentally treat procedures as unordered bags of words, discarding temporal structure necessary for cross-context transfer. Corpus scale delivers far larger gains than representation enrichment, revealing an architectural ceiling in current encoders. Our benchmark offers the first diagnostic framework separating genuine procedural understanding from surface-level memorization and gives tools for developing retrieval systems capable of dependable generalization. Resources available at our GitHub repository (https://github.com/qpiai/Proced_mem_bench).", "published": "2025-11-21", "updated": "2025-11-21", "authors": [ "Ishant Kohar", "Aswanth Krishnan" ], "categories": [ "cs.CL", "cs.AI", "cs.LG" ], "url": "https://arxiv.org/abs/2511.21730", "pdf_url": "https://arxiv.org/pdf/2511.21730", "matched_queries": [ "procedural-memory" ] }, { "arxiv_id": "2511.18112", "title": "EchoVLA: Synergistic Declarative Memory for VLA-Driven Mobile Manipulation", "abstract": "Recent progress in Vision-Language-Action (VLA) models has enabled embodied agents to interpret multimodal instructions and perform complex tasks. However, existing VLAs are mostly confined to short-horizon, table-top manipulation, lacking the memory and reasoning capability required for mobile manipulation, where agents must coordinate navigation and manipulation under changing spatial contexts. In this work, we present EchoVLA, a memory-aware VLA model for mobile manipulation. EchoVLA incorporates a synergistic declarative memory inspired by the human brain, consisting of a scene memory that maintains a collection of spatial-semantic maps and an episodic memory that stores task-level experiences with multimodal contextual features. The two memories are individually stored, updated, and retrieved based on current observations, task history, and instructions, and their retrieved representations are fused via coarse- and fine-grained attention to guide base-arm diffusion policies. To support large-scale training, we further introduce MoMani, an automated benchmark that generates expert-level trajectories through multimodal large language model (MLLM)-guided planning and feedback-driven refinement, supplemented with real-robot demonstrations. Comprehensive simulated and real-world results demonstrate that EchoVLA substantially improves overall performance, e.g., it achieves the highest success rates of 0.52 on manipulation/navigation tasks and 0.31 on mobile manipulation tasks in simulation, exceeding the strong baseline $π_{0.5}$ by +0.20 and +0.11, respectively.", "published": "2025-11-22", "updated": "2026-03-06", "authors": [ "Min Lin", "Xiwen Liang", "Bingqian Lin", "Liu Jingzhi", "Zijian Jiao", "Kehan Li", "Yu Sun", "Weijia Liufu", "Yuhan Ma", "Yuecheng Liu", "Shen Zhao", "Yuzheng Zhuang", "Xiaodan Liang" ], "categories": [ "cs.RO" ], "url": "https://arxiv.org/abs/2511.18112", "pdf_url": "https://arxiv.org/pdf/2511.18112", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2511.18423", "title": "General Agentic Memory Via Deep Research", "abstract": "Memory is critical for AI agents, yet the widely-adopted static memory, aiming to create readily available memory in advance, is inevitably subject to severe information loss. To address this limitation, we propose a novel framework called \\textbf{general agentic memory (GAM)}. GAM follows the principle of \"\\textbf{just-in time (JIT) compilation}\" where it focuses on creating optimized contexts for its client at runtime while keeping only simple but useful memory during the offline stage. To this end, GAM employs a duo-design with the following components. 1) \\textbf{Memorizer}, which highlights key historical information using a lightweight memory, while maintaining complete historical information within a universal page-store. 2) \\textbf{Researcher}, which retrieves and integrates useful information from the page-store for its online request guided by the pre-constructed memory. This design allows GAM to effectively leverage the agentic capabilities and test-time scalability of frontier large language models (LLMs), while also facilitating end-to-end performance optimization through reinforcement learning. In our experimental study, we demonstrate that GAM achieves substantial improvement on various memory-grounded task completion scenarios against existing memory systems.", "published": "2025-11-23", "updated": "2025-11-23", "authors": [ "B. Y. Yan", "Chaofan Li", "Hongjin Qian", "Shuqi Lu", "Zheng Liu" ], "categories": [ "cs.CL", "cs.AI", "cs.IR", "cs.LG" ], "url": "https://arxiv.org/abs/2511.18423", "pdf_url": "https://arxiv.org/pdf/2511.18423", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2511.19192", "title": "AME: An Efficient Heterogeneous Agentic Memory Engine for Smartphones", "abstract": "On-device agents on smartphones increasingly require continuously evolving memory to support personalized, context-aware, and long-term behaviors. To meet both privacy and responsiveness demands, user data is embedded as vectors and stored in a vector database for fast similarity search. However, most existing vector databases target server-class environments. When ported directly to smartphones, two gaps emerge: (G1) a mismatch between mobile SoC constraints and vector-database assumptions, including tight bandwidth budgets, limited on-chip memory, and stricter data type and layout constraints; and (G2) a workload mismatch, because on-device usage resembles a continuously learning memory, in which queries must coexist with frequent inserts, deletions, and ongoing index maintenance. To address these challenges, we propose AME, an on-device Agentic Memory Engine co-designed with modern smartphone SoCs. AME introduces two key techniques: (1) a hardware-aware, high-efficiency matrix pipeline that maximizes compute-unit utilization and exploits multi-level on-chip storage to sustain high throughput; and (2) a hardware- and workload-aware scheduling scheme that coordinates querying, insertion, and index rebuilding to minimize latency. We implement AME on Snapdragon 8-series SoCs and evaluate it on HotpotQA. In our experiments, AME improves query throughput by up to 1.4x at matched recall, achieves up to 7x faster index construction, and delivers up to 6x higher insertion throughput under concurrent query workloads.", "published": "2025-11-24", "updated": "2025-11-24", "authors": [ "Xinkui Zhao", "Qingyu Ma", "Yifan Zhang", "Hengxuan Lou", "Guanjie Cheng", "Shuiguang Deng", "Jianwei Yin" ], "categories": [ "cs.DC" ], "url": "https://arxiv.org/abs/2511.19192", "pdf_url": "https://arxiv.org/pdf/2511.19192", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2511.20297", "title": "Improving Language Agents through BREW", "abstract": "Large Language Model (LLM)-based agents are increasingly applied to tasks requiring structured reasoning, tool use, and environmental adaptation, such as data manipulation, multistep planning, and computer-use automation. However, despite their versatility, current training paradigms for model weight optimization methods, like PPO and GRPO, remain relatively impractical with their high computational overhead for rollout convergence. In addition, the resulting agent policies are difficult to interpret, adapt, or incrementally improve. To address this, we investigate creating and refining structured memory of experiential learning of an agent from its environment as an alternative route to agent optimization. We introduce BREW (Bootstrapping expeRientially-learned Environmental knoWledge), a framework for agent optimization for downstream tasks via KB construction and refinement. In our formulation, we introduce an effective method for partitioning agent memory for more efficient retrieval and refinement. BREW uses task graders and behavior rubrics to learn insights while leveraging state-space search for ensuring robustness from the noise and non-specificity in natural language. Empirical results on real world, domain-grounded benchmarks -- OSWorld, $τ^2$Bench, and SpreadsheetBench -- show BREW achieves $10-20\\%$ improvement in task precision, $10-15\\%$ reduction in API/tool calls leading to faster execution time, all while maintaining computational efficiency on par with base models. Unlike prior work where memory is treated as static context, we establish the KB as a modular and controllable substrate for agent optimization -- an explicit lever for shaping behavior in a transparent, interpretable, and extensible manner.", "published": "2025-11-25", "updated": "2025-11-25", "authors": [ "Shashank Kirtania", "Param Biyani", "Priyanshu Gupta", "Yasharth Bajpai", "Roshni Iyer", "Sumit Gulwani", "Gustavo Soares" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2511.20297", "pdf_url": "https://arxiv.org/pdf/2511.20297", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2511.20857", "title": "Evo-Memory: Benchmarking LLM Agent Test-time Learning with Self-Evolving Memory", "abstract": "Statefulness is essential for large language model (LLM) agents to perform long-term planning and problem-solving. This makes memory a critical component, yet its management and evolution remain largely underexplored. Existing evaluations mostly focus on static conversational settings, where memory is passively retrieved from dialogue to answer queries, overlooking the dynamic ability to accumulate and reuse experience across evolving task streams. In real-world environments such as interactive problem assistants or embodied agents, LLMs are required to handle continuous task streams, yet often fail to learn from accumulated interactions, losing valuable contextual insights, a limitation that calls for test-time evolution, where LLMs retrieve, integrate, and update memory continuously during deployment. To bridge this gap, we introduce Evo-Memory, a comprehensive streaming benchmark and framework for evaluating self-evolving memory in LLM agents. Evo-Memory structures datasets into sequential task streams, requiring LLMs to search, adapt, and evolve memory after each interaction. We unify and implement over ten representative memory modules and evaluate them across 10 diverse multi-turn goal-oriented and single-turn reasoning and QA datasets. To better benchmark experience reuse, we provide a baseline method, ExpRAG, for retrieving and utilizing prior experience, and further propose ReMem, an action-think-memory refine pipeline that tightly integrates reasoning, task actions, and memory updates to achieve continual improvement.", "published": "2025-11-25", "updated": "2026-05-18", "authors": [ "Tianxin Wei", "Noveen Sachdeva", "Benjamin Coleman", "Zhankui He", "Yuanchen Bei", "Xuying Ning", "Mengting Ai", "Yunzhe Li", "Jingrui He", "Ed H. Chi", "Chi Wang", "Shuo Chen", "Fernando Pereira", "Wang-Cheng Kang", "Derek Zhiyuan Cheng" ], "categories": [ "cs.CL", "cs.AI" ], "url": "https://arxiv.org/abs/2511.20857", "pdf_url": "https://arxiv.org/pdf/2511.20857", "matched_queries": [ "memory-evaluation" ] }, { "arxiv_id": "2512.00742", "title": "On the Regulatory Potential of User Interfaces for AI Agent Governance", "abstract": "AI agents that take actions in their environment autonomously over extended time horizons require robust governance interventions to curb their potentially consequential risks. Prior proposals for governing AI agents primarily target system-level safeguards (e.g., prompt injection monitors) or agent infrastructure (e.g., agent IDs). In this work, we explore a complementary approach: regulating user interfaces of AI agents as a way of enforcing transparency and behavioral requirements that then demand changes at the system and/or infrastructure levels. Specifically, we analyze 22 existing agentic systems to identify UI elements that play key roles in human-agent interaction and communication. We then synthesize those elements into six high-level interaction design patterns that hold regulatory potential (e.g., requiring agent memory to be editable). We conclude with policy recommendations based on our analysis. Our work exposes a new surface for regulatory action that supplements previous proposals for practical AI agent governance.", "published": "2025-11-30", "updated": "2025-11-30", "authors": [ "K. J. Kevin Feng", "Tae Soo Kim", "Rock Yuren Pang", "Faria Huq", "Tal August", "Amy X. Zhang" ], "categories": [ "cs.CY", "cs.AI" ], "url": "https://arxiv.org/abs/2512.00742", "pdf_url": "https://arxiv.org/pdf/2512.00742", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2512.02227", "title": "Orchestration Framework for Financial Agents: From Algorithmic Trading to Agentic Trading", "abstract": "The financial market is a mission-critical playground for AI agents due to its temporal dynamics and low signal-to-noise ratio. Building an effective algorithmic trading system may require a professional team to develop and test over the years. In this paper, we propose an orchestration framework for financial agents, which aims to democratize financial intelligence to the general public. We map each component of the traditional algorithmic trading system to agents, including planner, orchestrator, alpha agents, risk agents, portfolio agents, backtest agents, execution agents, audit agents, and memory agent. We present two in-house trading examples. For the stock trading task (hourly data from 04/2024 to 12/2024), our approach achieved a return of $20.42\\%$, a Sharpe ratio of 2.63, and a maximum drawdown of $-3.59\\%$, while the S&P 500 index yielded a return of $15.97\\%$. For the BTC trading task (minute data from 27/07/2025 to 13/08/2025), our approach achieved a return of $8.39\\%$, a Sharpe ratio of $0.38$, and a maximum drawdown of $-2.80\\%$, whereas the BTC price increased by $3.80\\%$. Our code is available on \\href{https://github.com/Open-Finance-Lab/AgenticTrading}{GitHub}.", "published": "2025-12-01", "updated": "2025-12-01", "authors": [ "Jifeng Li", "Arnav Grover", "Abraham Alpuerto", "Yupeng Cao", "Xiao-Yang Liu" ], "categories": [ "cs.MA", "cs.AI", "cs.CE", "cs.LG" ], "url": "https://arxiv.org/abs/2512.02227", "pdf_url": "https://arxiv.org/pdf/2512.02227", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2512.02228", "title": "STRIDE: A Systematic Framework for Selecting AI Modalities -- Agentic AI, AI Assistants, or LLM Calls", "abstract": "The rapid shift from stateless large language models (LLMs) to autonomous, goal-driven agents raises a central question: When is agentic AI truly necessary? While agents enable multi-step reasoning, persistent memory, and tool orchestration, deploying them indiscriminately leads to higher cost, complexity, and risk. We present STRIDE (Systematic Task Reasoning Intelligence Deployment Evaluator), a framework that provides principled recommendations for selecting between three modalities: (i) direct LLM calls, (ii) guided AI assistants, and (iii) fully autonomous agentic AI. STRIDE integrates structured task decomposition, dynamism attribution, and self-reflection requirement analysis to produce an Agentic Suitability Score, ensuring that full agentic autonomy is reserved for tasks with inherent dynamism or evolving context. Evaluated across 30 real-world tasks spanning SRE, compliance, and enterprise automation, STRIDE achieved 92% accuracy in modality selection, reduced unnecessary agent deployments by 45%, and cut resource costs by 37%. Expert validation over six months in SRE and compliance domains confirmed its practical utility, with domain specialists agreeing that STRIDE effectively distinguishes between tasks requiring simple LLM calls, guided assistants, or full agentic autonomy. This work reframes agent adoption as a necessity-driven design decision, ensuring autonomy is applied only when its benefits justify the costs.", "published": "2025-12-01", "updated": "2025-12-01", "authors": [ "Shubhi Asthana", "Bing Zhang", "Chad DeLuca", "Ruchi Mahindru", "Hima Patel" ], "categories": [ "cs.AI", "cs.LG" ], "url": "https://arxiv.org/abs/2512.02228", "pdf_url": "https://arxiv.org/pdf/2512.02228", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2512.02425", "title": "WorldMM: Dynamic Multimodal Memory Agent for Long Video Reasoning", "abstract": "Recent advances in video large language models have demonstrated strong capabilities in understanding short clips. However, scaling them to hours- or days-long videos remains highly challenging due to limited context capacity and the loss of critical visual details during abstraction. Existing memory-augmented methods mitigate this by leveraging textual summaries of video segments, yet they heavily rely on text and fail to utilize visual evidence when reasoning over complex scenes. Moreover, retrieving from fixed temporal scales further limits their flexibility in capturing events that span variable durations. To address this, we introduce WorldMM, a novel multimodal memory agent that constructs and retrieves from multiple complementary memories, encompassing both textual and visual representations. WorldMM comprises three types of memory: episodic memory indexes factual events across multiple temporal scales, semantic memory continuously updates high-level conceptual knowledge, and visual memory preserves detailed information about scenes. During inference, an adaptive retrieval agent iteratively selects the most relevant memory source and leverages multiple temporal granularities based on the query, continuing until it determines that sufficient information has been gathered. WorldMM significantly outperforms existing baselines across five long video question-answering benchmarks, achieving an average 8.4% performance gain over previous state-of-the-art methods, showing its effectiveness on long video reasoning.", "published": "2025-12-02", "updated": "2026-03-27", "authors": [ "Woongyeong Yeo", "Kangsan Kim", "Jaehong Yoon", "Sung Ju Hwang" ], "categories": [ "cs.CV", "cs.AI", "cs.CL", "cs.IR", "cs.LG" ], "url": "https://arxiv.org/abs/2512.02425", "pdf_url": "https://arxiv.org/pdf/2512.02425", "matched_queries": [ "agent-memory", "episodic-memory" ] }, { "arxiv_id": "2512.02458", "title": "Vision to Geometry: 3D Spatial Memory for Sequential Embodied MLLM Reasoning and Exploration", "abstract": "Embodied agents are expected to assist humans by actively exploring unknown environments and reasoning about spatial contexts. When deployed in real life, agents often face sequential tasks where each new task follows the completion of the previous one and may include infeasible objectives, such as searching for non-existent objects. However, most existing research focuses on isolated goals, overlooking the core challenge of sequential tasks: the ability to reuse spatial knowledge accumulated from previous explorations to guide subsequent reasoning and exploration. In this work, we investigate this underexplored yet practically significant embodied AI challenge. Specifically, we propose 3DSPMR, a 3D SPatial Memory Reasoning framework that utilizes Field-of-View (FoV) coverage as an explicit geometric prior. By integrating FoV-based constraints, 3DSPMR significantly enhances an agent's memory, reasoning, and exploration capabilities across sequential tasks. To facilitate research in this area, we further introduce SEER-Bench, a novel Sequential Embodied Exploration and Reasoning Benchmark that spans two foundational tasks: Embodied Question Answering (EQA) and Embodied Multi-modal Navigation (EMN). SEER-Bench uniquely incorporates both feasible and infeasible tasks to provide a rigorous and comprehensive evaluation of agent performance. Extensive experiments verify that 3DSPMR achieves substantial performance gains on both sequential EQA and EMN tasks.", "published": "2025-12-02", "updated": "2026-03-18", "authors": [ "Zhongyi Cai", "Yi Du", "Chen Wang", "Yu Kong" ], "categories": [ "cs.CV" ], "url": "https://arxiv.org/abs/2512.02458", "pdf_url": "https://arxiv.org/pdf/2512.02458", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2512.03627", "title": "MemVerse: Multimodal Memory for Lifelong Learning Agents", "abstract": "Despite rapid progress in large-scale language and vision models, AI agents still suffer from a fundamental limitation: they cannot remember. Without reliable memory, agents catastrophically forget past experiences, struggle with long-horizon reasoning, and fail to operate coherently in multimodal or interactive environments. We introduce MemVerse, a model-agnostic, plug-and-play memory framework that bridges fast parametric recall with hierarchical retrieval-based memory, enabling scalable and adaptive multimodal intelligence. MemVerse maintains short-term memory for recent context while transforming raw multimodal experiences into structured long-term memories organized as hierarchical knowledge graphs. This design supports continual consolidation, adaptive forgetting, and bounded memory growth. To handle real-time demands, MemVerse introduces a periodic distillation mechanism that compresses essential knowledge from long-term memory into the parametric model, allowing fast, differentiable recall while preserving interpretability. Extensive experiments demonstrate that MemVerse significantly improves multimodal reasoning and continual learning efficiency, empowering agents to remember, adapt, and reason coherently across extended interactions.", "published": "2025-12-03", "updated": "2026-06-02", "authors": [ "Junming Liu", "Yifei Sun", "Weihua Cheng", "Haodong Lei", "Yirong Chen", "Licheng Wen", "Xuemeng Yang", "Daocheng Fu", "Pinlong Cai", "Nianchen Deng", "Yi Yu", "Shuyue Hu", "Botian Shi", "Ding Wang" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2512.03627", "pdf_url": "https://arxiv.org/pdf/2512.03627", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2512.04668", "title": "Topology Matters: Measuring Memory Leakage in Multi-Agent LLMs", "abstract": "Graph topology is a fundamental determinant of memory leakage in multi-agent LLM systems, yet its effects remain poorly quantified. We introduce MAMA (Multi-Agent Memory Attack), a controlled evaluation framework for comparing topology-conditioned memory leakage in multi-agent LLM systems. MAMA operates on synthetic documents containing labeled Personally Identifiable Information (PII) entities, from which we generate sanitized task instructions. We execute a two-phase protocol: Engram (seeding private information into a target agent's memory) and Resonance (multi-round interaction where an attacker attempts extraction). Over 10 rounds, we measure leakage using a two-stage recovery criterion that combines exact-match extraction with LLM-based inference over the attacker's final output. We evaluate six canonical topologies (complete, circle, chain, tree, star, star-ring) across $n\\in\\{4,5,6\\}$, attacker-target placements, and base models. Results are consistent: denser connectivity, shorter attacker-target distance, and higher target centrality increase leakage; most leakage occurs in early rounds and then plateaus; model choice shifts absolute rates but preserves broad structural trends; spatiotemporal/location attributes leak more readily than identity credentials or regulated identifiers. We distill practical guidance for system design: favor sparse or hierarchical connectivity, maximize attacker-target separation, and restrict hub/shortcut pathways via topology-aware access control. Our code is available at https://github.com/llll121/mama-eval.", "published": "2025-12-04", "updated": "2026-06-03", "authors": [ "Jinbo Liu", "Defu Cao", "Yifei Wei", "Tianyao Su", "Yuan Liang", "Yushun Dong", "Yan Liu", "Yue Zhao", "Xiyang Hu" ], "categories": [ "cs.CR", "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2512.04668", "pdf_url": "https://arxiv.org/pdf/2512.04668", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2512.06688", "title": "PersonaMem-v2: Towards Personalized Intelligence via Learning Implicit User Personas and Agentic Memory", "abstract": "Personalization is one of the next milestones in advancing AI capability and alignment. We introduce PersonaMem-v2, the state-of-the-art dataset for LLM personalization that simulates 1,000 realistic user-chatbot interactions on 300+ scenarios, 20,000+ user preferences, and 128k-token context windows, where most user preferences are implicitly revealed to reflect real-world interactions. Using this data, we investigate how reinforcement fine-tuning enables a model to improve its long-context reasoning capabilities for user understanding and personalization. We also develop a framework for training an agentic memory system, which maintains a single, human-readable memory that grows with each user over time. In our experiments, frontier LLMs still struggle with implicit personalization, achieving only 37-48% accuracy. While they support long context windows, reasoning remains the bottleneck for implicit personalization tasks. Using reinforcement fine-tuning, we successfully train Qwen3-4B to outperforms GPT-5, reaching 53% accuracy in implicit personalization. Moreover, our agentic memory framework achieves state-of-the-art 55% accuracy while using 16x fewer input tokens, relying on a 2k-token memory instead of full 32k conversation histories. These results underscore the impact of our dataset and demonstrate agentic memory as a scalable path toward real-world personalized intelligence.", "published": "2025-12-07", "updated": "2025-12-07", "authors": [ "Bowen Jiang", "Yuan Yuan", "Maohao Shen", "Zhuoqun Hao", "Zhangchen Xu", "Zichen Chen", "Ziyi Liu", "Anvesh Rao Vijjini", "Jiashu He", "Hanchao Yu", "Radha Poovendran", "Gregory Wornell", "Lyle Ungar", "Dan Roth", "Sihao Chen", "Camillo Jose Taylor" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2512.06688", "pdf_url": "https://arxiv.org/pdf/2512.06688", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2512.09458", "title": "Architectures for Building Agentic AI", "abstract": "This chapter argues that the reliability of agentic and generative AI is chiefly an architectural property. We define agentic systems as goal-directed, tool-using decision makers operating in closed loops, and show how reliability emerges from principled componentisation (goal manager, planner, tool-router, executor, memory, verifiers, safety monitor, telemetry), disciplined interfaces (schema-constrained, validated, least-privilege tool calls), and explicit control and assurance loops. Building on classical foundations, we propose a practical taxonomy-tool-using agents, memory-augmented agents, planning and self-improvement agents, multi-agent systems, and embodied or web agents - and analyse how each pattern reshapes the reliability envelope and failure modes. We distil design guidance on typed schemas, idempotency, permissioning, transactional semantics, memory provenance and hygiene, runtime governance (budgets, termination conditions), and simulate-before-actuate safeguards.", "published": "2025-12-10", "updated": "2025-12-10", "authors": [ "Sławomir Nowaczyk" ], "categories": [ "cs.AI", "cs.LG" ], "url": "https://arxiv.org/abs/2512.09458", "pdf_url": "https://arxiv.org/pdf/2512.09458", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2512.10166", "title": "Emergent Collective Memory in Decentralized Multi-Agent AI Systems", "abstract": "We demonstrate how collective memory emerges in decentralized multi-agent systems through the interplay between individual agent memory and environmental trace communication. Our agents maintain internal memory states while depositing persistent environmental traces, creating a spatially distributed collective memory without centralized control. Comprehensive validation across five environmental conditions (20x20 to 50x50 grids, 5-20 agents, 50 runs per configuration) reveals a critical asymmetry: individual memory alone provides 68.7% performance improvement over no-memory baselines (1563.87 vs 927.23, p < 0.001), while environmental traces without memory fail completely. This demonstrates that memory functions independently but traces require cognitive infrastructure for interpretation. Systematic density-sweep experiments (rho in [0.049, 0.300], up to 625 agents) validate our theoretical phase transition prediction. On realistic large grids (30x30, 50x50), stigmergic coordination dominates above rho ~ 0.20, with traces outperforming memory by 36-41% on composite metrics despite lower food efficiency. The experimental crossover confirms the predicted critical density rho_c = 0.230 within 13% error.", "published": "2025-12-10", "updated": "2025-12-10", "authors": [ "Khushiyant" ], "categories": [ "cs.MA" ], "url": "https://arxiv.org/abs/2512.10166", "pdf_url": "https://arxiv.org/pdf/2512.10166", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2512.10696", "title": "Remember Me, Refine Me: A Dynamic Procedural Memory Framework for Experience-Driven Agent Evolution", "abstract": "Procedural memory enables large language model (LLM) agents to internalize \"how-to\" knowledge, theoretically reducing redundant trial-and-error. However, existing frameworks predominantly suffer from a \"passive accumulation\" paradigm, treating memory as a static append-only archive. To bridge the gap between static storage and dynamic reasoning, we propose $\\textbf{ReMe}$ ($\\textit{Remember Me, Refine Me}$), a comprehensive framework for experience-driven agent evolution. ReMe innovates across the memory lifecycle via three mechanisms: 1) $\\textit{multi-faceted distillation}$, which extracts fine-grained experiences by recognizing success patterns, analyzing failure triggers and generating comparative insights; 2) $\\textit{context-adaptive reuse}$, which tailors historical insights to new contexts via scenario-aware indexing; and 3) $\\textit{utility-based refinement}$, which autonomously adds valid memories and prunes outdated ones to maintain a compact, high-quality experience pool. Extensive experiments on BFCL-V3 and AppWorld demonstrate that ReMe establishes a new state-of-the-art in agent memory system. Crucially, we observe a significant memory-scaling effect: Qwen3-8B equipped with ReMe outperforms larger, memoryless Qwen3-14B, suggesting that self-evolving memory provides a computation-efficient pathway for lifelong learning. We release our code and the $\\texttt{reme.library}$ dataset to facilitate further research.", "published": "2025-12-11", "updated": "2026-04-15", "authors": [ "Zouying Cao", "Jiaji Deng", "Li Yu", "Weikang Zhou", "Zhaoyang Liu", "Bolin Ding", "Hai Zhao" ], "categories": [ "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2512.10696", "pdf_url": "https://arxiv.org/pdf/2512.10696", "matched_queries": [ "agent-memory", "procedural-memory" ] }, { "arxiv_id": "2512.11303", "title": "Unifying Dynamic Tool Creation and Cross-Task Experience Sharing through Cognitive Memory Architecture", "abstract": "Large Language Model agents face fundamental challenges in adapting to novel tasks due to limitations in tool availability and experience reuse. Existing approaches either rely on predefined tools with limited coverage or build tools from scratch without leveraging past experiences, leading to inefficient exploration and suboptimal performance. We introduce SMITH (Shared Memory Integrated Tool Hub), a unified cognitive architecture that seamlessly integrates dynamic tool creation with cross-task experience sharing through hierarchical memory organization. SMITH organizes agent memory into procedural, semantic, and episodic components, enabling systematic capability expansion while preserving successful execution patterns. Our approach formalizes tool creation as iterative code generation within controlled sandbox environments and experience sharing through episodic memory retrieval with semantic similarity matching. We further propose a curriculum learning strategy based on agent-ensemble difficulty re-estimation. Extensive experiments on the GAIA benchmark demonstrate SMITH's effectiveness, achieving 81.8% Pass@1 accuracy and outperforming state-of-the-art baselines including Alita (75.2%) and Memento (70.9%). Our work establishes a foundation for building truly adaptive agents that continuously evolve their capabilities through principled integration of tool creation and experience accumulation.", "published": "2025-12-12", "updated": "2025-12-12", "authors": [ "Jiarun Liu", "Shiyue Xu", "Yang Li", "Shangkun Liu", "Yongli Yu", "Peng Cao" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2512.11303", "pdf_url": "https://arxiv.org/pdf/2512.11303", "matched_queries": [ "agent-memory", "episodic-memory" ] }, { "arxiv_id": "2512.12686", "title": "Memoria: A Scalable Agentic Memory Framework for Personalized Conversational AI", "abstract": "Agentic memory is emerging as a key enabler for large language models (LLM) to maintain continuity, personalization, and long-term context in extended user interactions, critical capabilities for deploying LLMs as truly interactive and adaptive agents. Agentic memory refers to the memory that provides an LLM with agent-like persistence: the ability to retain and act upon information across conversations, similar to how a human would. We present Memoria, a modular memory framework that augments LLM-based conversational systems with persistent, interpretable, and context-rich memory. Memoria integrates two complementary components: dynamic session-level summarization and a weighted knowledge graph (KG)-based user modelling engine that incrementally captures user traits, preferences, and behavioral patterns as structured entities and relationships. This hybrid architecture enables both short-term dialogue coherence and long-term personalization while operating within the token constraints of modern LLMs. We demonstrate how Memoria enables scalable, personalized conversational artificial intelligence (AI) by bridging the gap between stateless LLM interfaces and agentic memory systems, offering a practical solution for industry applications requiring adaptive and evolving user experiences.", "published": "2025-12-14", "updated": "2025-12-14", "authors": [ "Samarth Sarin", "Lovepreet Singh", "Bhaskarjit Sarmah", "Dhagash Mehta" ], "categories": [ "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2512.12686", "pdf_url": "https://arxiv.org/pdf/2512.12686", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2512.12818", "title": "Hindsight is 20/20: Building Agent Memory that Retains, Recalls, and Reflects", "abstract": "Agent memory has been touted as a dimension of growth for LLM-based applications, enabling agents that can accumulate experience, adapt across sessions, and move beyond single-shot question answering. The current generation of agent memory systems treats memory as an external layer that extracts salient snippets from conversations, stores them in vector or graph-based stores, and retrieves top-k items into the prompt of an otherwise stateless model. While these systems improve personalization and context carry-over, they still blur the line between evidence and inference, struggle to organize information over long horizons, and offer limited support for agents that must explain their reasoning. We present Hindsight, a memory architecture that treats agent memory as a structured, first-class substrate for reasoning by organizing it into four logical networks that distinguish world facts, agent experiences, synthesized entity summaries, and evolving beliefs. This framework supports three core operations -- retain, recall, and reflect -- that govern how information is added, accessed, and updated. Under this abstraction, a temporal, entity aware memory layer incrementally turns conversational streams into a structured, queryable memory bank, while a reflection layer reasons over this bank to produce answers and to update information in a traceable way. On key long-horizon conversational memory benchmarks like LongMemEval and LoCoMo, Hindsight with an open-source 20B model lifts overall accuracy from 39% to 83.6% over a full-context baseline with the same backbone and outperforms full context GPT-4o. Scaling the backbone further pushes Hindsight to 91.4% on LongMemEval and up to 89.61% on LoCoMo (vs. 75.78% for the strongest prior open system), consistently outperforming existing memory architectures on multi-session and open-domain questions.", "published": "2025-12-14", "updated": "2025-12-14", "authors": [ "Chris Latimer", "Nicoló Boschi", "Andrew Neeser", "Chris Bartholomew", "Gaurav Srivastava", "Xuan Wang", "Naren Ramakrishnan" ], "categories": [ "cs.CL", "cs.AI", "cs.IR", "cs.LG" ], "url": "https://arxiv.org/abs/2512.12818", "pdf_url": "https://arxiv.org/pdf/2512.12818", "matched_queries": [ "agent-memory", "memory-evaluation" ] }, { "arxiv_id": "2512.12856", "title": "Forgetful but Faithful: A Cognitive Memory Architecture and Benchmark for Privacy-Aware Generative Agents", "abstract": "As generative agents become increasingly sophisticated and deployed in long-term interactive scenarios, their memory management capabilities emerge as a critical bottleneck for both performance and privacy. Current approaches either maintain unlimited memory stores, leading to computational intractability and privacy concerns, or employ simplistic forgetting mechanisms that compromise agent coherence and functionality. This paper introduces the Memory-Aware Retention Schema (MaRS), a novel framework for human-centered memory management in generative agents, coupled with six theoretically-grounded forgetting policies that balance performance, privacy, and computational efficiency. We present the Forgetful but Faithful Agent (FiFA) benchmark, a comprehensive evaluation framework that assesses agent performance across narrative coherence, goal completion, social recall accuracy, privacy preservation, and cost efficiency. Through extensive experimentation involving 300 evaluation runs across multiple memory budgets and agent configurations, we demonstrate that our hybrid forgetting policy achieves superior performance (composite score: 0.911) while maintaining computational tractability and privacy guarantees. Our work establishes new benchmarks for memory-budgeted agent evaluation and provides practical guidelines for deploying generative agents in resource-constrained, privacy-sensitive environments. The theoretical foundations, implementation framework, and empirical results contribute to the emerging field of human-centered AI by addressing fundamental challenges in agent memory management that directly impact user trust, system scalability, and regulatory compliance.", "published": "2025-12-14", "updated": "2025-12-14", "authors": [ "Saad Alqithami" ], "categories": [ "cs.AI", "cs.LG" ], "url": "https://arxiv.org/abs/2512.12856", "pdf_url": "https://arxiv.org/pdf/2512.12856", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2512.12967", "title": "QwenLong-L1.5: Post-Training Recipe for Long-Context Reasoning and Memory Management", "abstract": "We introduce QwenLong-L1.5, a model that achieves superior long-context reasoning capabilities through systematic post-training innovations. The key technical breakthroughs of QwenLong-L1.5 are as follows: (1) Long-Context Data Synthesis Pipeline: We develop a systematic synthesis framework that generates challenging reasoning tasks requiring multi-hop grounding over globally distributed evidence. By deconstructing documents into atomic facts and their underlying relationships, and then programmatically composing verifiable reasoning questions, our approach creates high-quality training data at scale, moving substantially beyond simple retrieval tasks to enable genuine long-range reasoning capabilities. (2) Stabilized Reinforcement Learning for Long-Context Training: To overcome the critical instability in long-context RL, we introduce task-balanced sampling with task-specific advantage estimation to mitigate reward bias, and propose Adaptive Entropy-Controlled Policy Optimization (AEPO) that dynamically regulates exploration-exploitation trade-offs. (3) Memory-Augmented Architecture for Ultra-Long Contexts: Recognizing that even extended context windows cannot accommodate arbitrarily long sequences, we develop a memory management framework with multi-stage fusion RL training that seamlessly integrates single-pass reasoning with iterative memory-based processing for tasks exceeding 4M tokens. Based on Qwen3-30B-A3B-Thinking, QwenLong-L1.5 achieves performance comparable to GPT-5 and Gemini-2.5-Pro on long-context reasoning benchmarks, surpassing its baseline by 9.90 points on average. On ultra-long tasks (1M~4M tokens), QwenLong-L1.5's memory-agent framework yields a 9.48-point gain over the agent baseline. Additionally, the acquired long-context reasoning ability translates to enhanced performance in general domains like scientific reasoning, memory tool using, and extended dialogue.", "published": "2025-12-15", "updated": "2025-12-15", "authors": [ "Weizhou Shen", "Ziyi Yang", "Chenliang Li", "Zhiyuan Lu", "Miao Peng", "Huashan Sun", "Yingcheng Shi", "Shengyi Liao", "Shaopeng Lai", "Bo Zhang", "Dayiheng Liu", "Fei Huang", "Jingren Zhou", "Ming Yan" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2512.12967", "pdf_url": "https://arxiv.org/pdf/2512.12967", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2512.13564", "title": "Memory in the Age of AI Agents", "abstract": "Memory has emerged, and will continue to remain, a core capability of foundation model-based agents. As research on agent memory rapidly expands and attracts unprecedented attention, the field has also become increasingly fragmented. Existing works that fall under the umbrella of agent memory often differ substantially in their motivations, implementations, and evaluation protocols, while the proliferation of loosely defined memory terminologies has further obscured conceptual clarity. Traditional taxonomies such as long/short-term memory have proven insufficient to capture the diversity of contemporary agent memory systems. This work aims to provide an up-to-date landscape of current agent memory research. We begin by clearly delineating the scope of agent memory and distinguishing it from related concepts such as LLM memory, retrieval augmented generation (RAG), and context engineering. We then examine agent memory through the unified lenses of forms, functions, and dynamics. From the perspective of forms, we identify three dominant realizations of agent memory, namely token-level, parametric, and latent memory. From the perspective of functions, we propose a finer-grained taxonomy that distinguishes factual, experiential, and working memory. From the perspective of dynamics, we analyze how memory is formed, evolved, and retrieved over time. To support practical development, we compile a comprehensive summary of memory benchmarks and open-source frameworks. Beyond consolidation, we articulate a forward-looking perspective on emerging research frontiers, including memory automation, reinforcement learning integration, multimodal memory, multi-agent memory, and trustworthiness issues. We hope this survey serves not only as a reference for existing work, but also as a conceptual foundation for rethinking memory as a first-class primitive in the design of future agentic intelligence.", "published": "2025-12-15", "updated": "2026-01-13", "authors": [ "Yuyang Hu", "Shichun Liu", "Yanwei Yue", "Guibin Zhang", "Boyang Liu", "Fangyi Zhu", "Jiahang Lin", "Honglin Guo", "Shihan Dou", "Zhiheng Xi", "Senjie Jin", "Jiejun Tan", "Yanbin Yin", "Jiongnan Liu", "Zeyu Zhang", "Zhongxiang Sun", "Yutao Zhu", "Hao Sun", "Boci Peng", "Zhenrong Cheng", "Xuanbo Fan", "Jiaxin Guo", "Xinlei Yu", "Zhenhong Zhou", "Zewen Hu", "Jiahao Huo", "Junhao Wang", "Yuwei Niu", "Yu Wang", "Zhenfei Yin", "Xiaobin Hu", "Yue Liao", "Qiankun Li", "Kun Wang", "Wangchunshu Zhou", "Yixin Liu", "Dawei Cheng", "Qi Zhang", "Tao Gui", "Shirui Pan", "Yan Zhang", "Philip Torr", "Zhicheng Dou", "Ji-Rong Wen", "Xuanjing Huang", "Yu-Gang Jiang", "Shuicheng Yan" ], "categories": [ "cs.CL", "cs.AI" ], "url": "https://arxiv.org/abs/2512.13564", "pdf_url": "https://arxiv.org/pdf/2512.13564", "matched_queries": [ "agent-memory", "memory-evaluation" ] }, { "arxiv_id": "2512.16301", "title": "Adaptation of Agentic AI: A Survey of Post-Training, Memory, and Skills", "abstract": "Large language model (LLM) agents are moving beyond prompting alone. ChatGPT marked the rise of general-purpose LLM assistants, DeepSeek showed that on-policy reinforcement learning with verifiable rewards can improve reasoning and tool use, and OpenClaw highlights a newer direction in which agents accumulate persistent memory and reusable skills. Yet the research landscape remains fragmented across post-training, retrieval, memory, and skill systems. This survey studies these developments under a single notion of \\emph{adaptation}: improving an agent, its tools, or their interaction after pretraining. We organize the field with a four-paradigm framework spanning agent adaptation and tool adaptation. On the agent side, A1 (tool-execution-signaled) and A2 (agent-output-signaled) improve the agent itself through supervised fine-tuning, preference optimization, and reinforcement learning with verifiable rewards. On the tool side, T1 (agent-agnostic) provides reusable pre-trained modules any agent can call, while T2 (agent-supervised) uses the agent's outputs to train memory systems, skill libraries, or lightweight subagents. Using this framework, we review post-training methods, adaptive memory architectures, and agent skills; compare their trade-offs in cost, flexibility, and generalization; and summarize evaluation practices across deep research, software development, computer use, and drug discovery. We conclude by outlining open problems in agent-tool co-adaptation, continual learning, safety, and efficient deployment.", "published": "2025-12-18", "updated": "2026-03-09", "authors": [ "Pengcheng Jiang", "Jiacheng Lin", "Zhiyi Shi", "Zifeng Wang", "Luxi He", "Yichen Wu", "Ming Zhong", "Peiyang Song", "Qizheng Zhang", "Heng Wang", "Xueqiang Xu", "Hanwen Xu", "Pengrui Han", "Dylan Zhang", "Jiashuo Sun", "Chaoqi Yang", "Kun Qian", "Tian Wang", "Changran Hu", "Manling Li", "Quanzheng Li", "Hao Peng", "Sheng Wang", "Jingbo Shang", "Chao Zhang", "Jiaxuan You", "Liyuan Liu", "Pan Lu", "Yu Zhang", "Heng Ji", "Yejin Choi", "Dawn Song", "Jimeng Sun", "Jiawei Han" ], "categories": [ "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2512.16301", "pdf_url": "https://arxiv.org/pdf/2512.16301", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2512.16962", "title": "MemoryGraft: Persistent Compromise of LLM Agents via Poisoned Experience Retrieval", "abstract": "Large Language Model (LLM) agents increasingly rely on long-term memory and Retrieval-Augmented Generation (RAG) to persist experiences and refine future performance. While this experience learning capability enhances agentic autonomy, it introduces a critical, unexplored attack surface, i.e., the trust boundary between an agent's reasoning core and its own past. In this paper, we introduce MemoryGraft. It is a novel indirect injection attack that compromises agent behavior not through immediate jailbreaks, but by implanting malicious successful experiences into the agent's long-term memory. Unlike traditional prompt injections that are transient, or standard RAG poisoning that targets factual knowledge, MemoryGraft exploits the agent's semantic imitation heuristic which is the tendency to replicate patterns from retrieved successful tasks. We demonstrate that an attacker who can supply benign ingestion-level artifacts that the agent reads during execution can induce it to construct a poisoned RAG store where a small set of malicious procedure templates is persisted alongside benign experiences. When the agent later encounters semantically similar tasks, union retrieval over lexical and embedding similarity reliably surfaces these grafted memories, and the agent adopts the embedded unsafe patterns, leading to persistent behavioral drift across sessions. We validate MemoryGraft on MetaGPT's DataInterpreter agent with GPT-4o and find that a small number of poisoned records can account for a large fraction of retrieved experiences on benign workloads, turning experience-based self-improvement into a vector for stealthy and durable compromise. To facilitate reproducibility and future research, our code and evaluation data are available at https://github.com/Jacobhhy/Agent-Memory-Poisoning.", "published": "2025-12-18", "updated": "2025-12-18", "authors": [ "Saksham Sahai Srivastava", "Haoyu He" ], "categories": [ "cs.CR", "cs.AI", "cs.LG" ], "url": "https://arxiv.org/abs/2512.16962", "pdf_url": "https://arxiv.org/pdf/2512.16962", "matched_queries": [ "agent-memory", "long-term-memory" ] }, { "arxiv_id": "2512.18337", "title": "Towards Efficient Agents: A Co-Design of Inference Architecture and System", "abstract": "The rapid development of large language model (LLM)-based agents has unlocked new possibilities for autonomous multi-turn reasoning and tool-augmented decision-making. However, their real-world deployment is hindered by severe inefficiencies that arise not from isolated model inference, but from the systemic latency accumulated across reasoning loops, context growth, and heterogeneous tool interactions. This paper presents AgentInfer, a unified framework for end-to-end agent acceleration that bridges inference optimization and architectural design. We decompose the problem into four synergistic components: AgentCollab, a hierarchical dual-model reasoning framework that balances large- and small-model usage through dynamic role assignment; AgentSched, a cache-aware hybrid scheduler that minimizes latency under heterogeneous request patterns; AgentSAM, a suffix-automaton-based speculative decoding method that reuses multi-session semantic memory to achieve low-overhead inference acceleration; and AgentCompress, a semantic compression mechanism that asynchronously distills and reorganizes agent memory without disrupting ongoing reasoning. Together, these modules form a Self-Evolution Engine capable of sustaining efficiency and cognitive stability throughout long-horizon reasoning tasks. Experiments on the BrowseComp-zh and DeepDiver benchmarks demonstrate that through the synergistic collaboration of these methods, AgentInfer reduces ineffective token consumption by over 50%, achieving an overall 1.8-2.5 times speedup with preserved accuracy. These results underscore that optimizing for agentic task completion-rather than merely per-token throughput-is the key to building scalable, efficient, and self-improving intelligent systems.", "published": "2025-12-20", "updated": "2026-02-24", "authors": [ "Weizhe Lin", "Hui-Ling Zhen", "Shuai Yang", "Xian Wang", "Renxi Liu", "Hanting Chen", "Wangze Zhang", "Chuansai Zhou", "Yiming Li", "Chen Chen", "Xing Li", "Zhiyuan Yang", "Xiaosong Li", "Xianzhi Yu", "Zhenhua Dong", "Mingxuan Yuan", "Yunhe Wang" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2512.18337", "pdf_url": "https://arxiv.org/pdf/2512.18337", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2512.18571", "title": "ESearch-R1: Learning Cost-Aware MLLM Agents for Interactive Embodied Search via Reinforcement Learning", "abstract": "Multimodal Large Language Models (MLLMs) have empowered embodied agents with remarkable capabilities in planning and reasoning. However, when facing ambiguous natural language instructions (e.g., \"fetch the tool\" in a cluttered room), current agents often fail to balance the high cost of physical exploration against the cognitive cost of human interaction. They typically treat disambiguation as a passive perception problem, lacking the strategic reasoning to minimize total task execution costs. To bridge this gap, we propose ESearch-R1, a cost-aware embodied reasoning framework that unifies interactive dialogue (Ask), episodic memory retrieval (GetMemory), and physical navigation (Navigate) into a single decision process. We introduce HC-GRPO (Heterogeneous Cost-Aware Group Relative Policy Optimization). Unlike traditional PPO which relies on a separate value critic, HC-GRPO optimizes the MLLM by sampling groups of reasoning trajectories and reinforcing those that achieve the optimal trade-off between information gain and heterogeneous costs (e.g., navigate time, and human attention). Extensive experiments in AI2-THOR demonstrate that ESearch-R1 significantly outperforms standard ReAct-based agents. It improves task success rates while reducing total operational costs by approximately 50\\%, validating the effectiveness of GRPO in aligning MLLM agents with physical world constraints.", "published": "2025-12-21", "updated": "2025-12-21", "authors": [ "Weijie Zhou", "Xuangtang Xiong", "Ye Tian", "Lijun Yue", "Xinyu Wu", "Wei Li", "Chaoyang Zhao", "Honghui Dong", "Ming Tang", "Jinqiao Wang", "Zhengyou Zhang" ], "categories": [ "cs.AI", "cs.CV" ], "url": "https://arxiv.org/abs/2512.18571", "pdf_url": "https://arxiv.org/pdf/2512.18571", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2512.18746", "title": "MemEvolve: Meta-Evolution of Agent Memory Systems", "abstract": "Self-evolving memory systems are unprecedentedly reshaping the evolutionary paradigm of large language model (LLM)-based agents. Prior work has predominantly relied on manually engineered memory architectures to store trajectories, distill experience, and synthesize reusable tools, enabling agents to evolve on the fly within environment interactions. However, this paradigm is fundamentally constrained by the staticity of the memory system itself: while memory facilitates agent-level evolving, the underlying memory architecture cannot be meta-adapted to diverse task contexts. To address this gap, we propose MemEvolve, a meta-evolutionary framework that jointly evolves agents' experiential knowledge and their memory architecture, allowing agent systems not only to accumulate experience but also to progressively refine how they learn from it. To ground MemEvolve in prior research and foster openness in future self-evolving systems, we introduce EvolveLab, a unified self-evolving memory codebase that distills twelve representative memory systems into a modular design space (encode, store, retrieve, manage), providing both a standardized implementation substrate and a fair experimental arena. Extensive evaluations on four challenging agentic benchmarks demonstrate that MemEvolve achieves (I) substantial performance gains, improving frameworks such as SmolAgent and Flash-Searcher by up to $17.06\\%$; and (II) strong cross-task and cross-LLM generalization, designing memory architectures that transfer effectively across diverse benchmarks and backbone models.", "published": "2025-12-21", "updated": "2025-12-21", "authors": [ "Guibin Zhang", "Haotian Ren", "Chong Zhan", "Zhenhong Zhou", "Junhao Wang", "He Zhu", "Wangchunshu Zhou", "Shuicheng Yan" ], "categories": [ "cs.CL", "cs.MA" ], "url": "https://arxiv.org/abs/2512.18746", "pdf_url": "https://arxiv.org/pdf/2512.18746", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2512.18950", "title": "Learning Hierarchical Procedural Memory for LLM Agents through Bayesian Selection and Contrastive Refinement", "abstract": "We present MACLA, a framework that decouples reasoning from learning by maintaining a frozen large language model while performing all adaptation in an external hierarchical procedural memory. MACLA extracts reusable procedures from trajectories, tracks reliability via Bayesian posteriors, selects actions through expected-utility scoring, and refines procedures by contrasting successes and failures. Across four benchmarks (ALFWorld, WebShop, TravelPlanner, InterCodeSQL), MACLA achieves 78.1 percent average performance, outperforming all baselines. On ALFWorld unseen tasks, MACLA reaches 90.3 percent with 3.1 percent positive generalization. The system constructs memory in 56 seconds, 2800 times faster than the state-of-the-art LLM parameter-training baseline, compressing 2851 trajectories into 187 procedures. Experimental results demonstrate that structured external memory with Bayesian selection and contrastive refinement enables sample-efficient, interpretable, and continually improving agents without LLM parameter updates.", "published": "2025-12-22", "updated": "2025-12-22", "authors": [ "Saman Forouzandeh", "Wei Peng", "Parham Moradi", "Xinghuo Yu", "Mahdi Jalili" ], "categories": [ "cs.LG", "cs.AI" ], "url": "https://arxiv.org/abs/2512.18950", "pdf_url": "https://arxiv.org/pdf/2512.18950", "matched_queries": [ "procedural-memory" ] }, { "arxiv_id": "2512.19537", "title": "Event Extraction in Large Language Model", "abstract": "Large language models (LLMs) and multimodal LLMs are changing event extraction (EE): prompting and generation can often produce structured outputs in zero shot or few shot settings. Yet LLM based pipelines face deployment gaps, including hallucinations under weak constraints, fragile temporal and causal linking over long contexts and across documents, and limited long horizon knowledge management within a bounded context window. We argue that EE should be viewed as a system component that provides a cognitive scaffold for LLM centered solutions. Event schemas and slot constraints create interfaces for grounding and verification; event centric structures act as controlled intermediate representations for stepwise reasoning; event links support relation aware retrieval with graph based RAG; and event stores offer updatable episodic and agent memory beyond the context window. This survey covers EE in text and multimodal settings, organizing tasks and taxonomy, tracing method evolution from rule based and neural models to instruction driven and generative frameworks, and summarizing formulations, decoding strategies, architectures, representations, datasets, and evaluation. We also review cross lingual, low resource, and domain specific settings, and highlight open challenges and future directions for reliable event centric systems. Finally, we outline open challenges and future directions that are central to the LLM era, aiming to evolve EE from static extraction into a structurally reliable, agent ready perception and memory layer for open world systems.", "published": "2025-12-22", "updated": "2025-12-22", "authors": [ "Bobo Li", "Xudong Han", "Jiang Liu", "Yuzhe Ding", "Liqiang Jing", "Zhaoqi Zhang", "Jinheng Li", "Xinya Du", "Fei Li", "Meishan Zhang", "Min Zhang", "Aixin Sun", "Philip S. Yu", "Hao Fei" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2512.19537", "pdf_url": "https://arxiv.org/pdf/2512.19537", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2512.20111", "title": "ABBEL: Learning Natural-Language Belief States for Memory-Efficient Interaction", "abstract": "As the time horizons of sequential decision-making tasks grow, keeping full interaction histories in model context becomes increasingly costly. Recent work reduces context lengths by instead conditioning decision-making agents on recursively updated natural-language summaries, which are concise and interpretable. However, they underperform agents with access to the full context, suggesting that they fail to generate sufficient summaries. To address this we propose ABBEL, a recursive summarization framework that isolates and directly supervises each summary's information contents in the form of explicit natural-language belief states. First, we analyze the belief states generated by frontier models under ABBEL across five domains, and verify that performance is often degraded due to omitting or incorrectly updating information. We also discover settings where models use memory inefficiently by retaining extraneous information. We target these limitations by fine-tuning with two RL-based methods: belief grading, which reduces update errors by rewarding belief generations based on their information content, and peak belief penalties, which encourage compressing the beliefs with the greatest memory footprints. We demonstrate that these methods significantly reduce the performance gap with full context models, and enable ABBEL to outperform prior memory agent work by 40% while using 67% of the memory. Our code is available at https://github.com/jakob-bjorner/optimal-explorer-dev", "published": "2025-12-23", "updated": "2026-06-04", "authors": [ "Aly Lidayan", "Jakob Bjorner", "Satvik Golechha", "Kartik Goyal", "Alane Suhr" ], "categories": [ "cs.CL", "cs.AI", "cs.LG" ], "url": "https://arxiv.org/abs/2512.20111", "pdf_url": "https://arxiv.org/pdf/2512.20111", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2512.21567", "title": "Beyond Heuristics: A Decision-Theoretic Framework for Agent Memory Management", "abstract": "External memory is a key component of modern large language model (LLM) systems, enabling long-term interaction and personalization. Despite its importance, memory management is still largely driven by hand-designed heuristics, offering little insight into the long-term and uncertain consequences of memory decisions. In practice, choices about what to read or write shape future retrieval and downstream behavior in ways that are difficult to anticipate. We argue that memory management should be viewed as a sequential decision-making problem under uncertainty, where the utility of memory is delayed and dependent on future interactions. To this end, we propose DAM (Decision-theoretic Agent Memory), a decision-theoretic framework that decomposes memory management into immediate information access and hierarchical storage maintenance. Within this architecture, candidate operations are evaluated via value functions and uncertainty estimators, enabling an aggregate policy to arbitrate decisions based on estimated long-term utility and risk. Our contribution is not a new algorithm, but a principled reframing that clarifies the limitations of heuristic approaches and provides a foundation for future research on uncertainty-aware memory systems.", "published": "2025-12-25", "updated": "2025-12-25", "authors": [ "Changzhi Sun", "Xiangyu Chen", "Jixiang Luo", "Dell Zhang", "Xuelong Li" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2512.21567", "pdf_url": "https://arxiv.org/pdf/2512.21567", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2512.22716", "title": "Memento 2: Learning by Stateful Reflective Memory", "abstract": "We present a theoretical study of continual and experiential learning in large language model agents that combine episodic memory with reinforcement learning. We argue that the key mechanism for continual adaptation, without updating model parameters, is reflection: the agent's ability to use past experience to guide future actions. Empirical findings suggest that episodic, experience-driven reflection enables generalised adaptation across a wide range of open-ended, long-horizon tasks. This indicates that efficient learning can occur during deployment and weakens the traditional separation between training and testing. Motivated by this, we introduce the Stateful Reflective Decision Process, a formal model of reflective memory dynamics. In this abstraction, an agent maintains an episodic memory and performs two core operations. Writing stores interaction outcomes and plays the role of policy evaluation. Reading retrieves relevant past cases to inform decisions and plays the role of policy improvement. This perspective treats reflective memory as a control object that can be analysed using classical reinforcement learning tools. We then develop a read-write reflective learning framework by integrating retrieval into soft policy iteration and establish convergence guarantees. We show that as memory grows and provides denser coverage of the state space, the resulting composite policy converges to the optimal solution. Overall, this framework connects practical memory-based methods with principled reinforcement learning, providing a rigorous mathematical basis for building reflective, memory-embedded agents capable of continual general-purpose learning.", "published": "2025-12-27", "updated": "2026-01-29", "authors": [ "Jun Wang" ], "categories": [ "cs.AI", "cs.CV", "cs.LG" ], "url": "https://arxiv.org/abs/2512.22716", "pdf_url": "https://arxiv.org/pdf/2512.22716", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2512.23343", "title": "AI Meets Brain: Memory Systems from Cognitive Neuroscience to Autonomous Agents", "abstract": "Memory serves as the pivotal nexus bridging past and future, providing both humans and AI systems with invaluable concepts and experience to navigate complex tasks. Recent research on autonomous agents has increasingly focused on designing efficient memory workflows by drawing on cognitive neuroscience. However, constrained by interdisciplinary barriers, existing works struggle to assimilate the essence of human memory mechanisms. To bridge this gap, we systematically synthesizes interdisciplinary knowledge of memory, connecting insights from cognitive neuroscience with LLM-driven agents. Specifically, we first elucidate the definition and function of memory along a progressive trajectory from cognitive neuroscience through LLMs to agents. We then provide a comparative analysis of memory taxonomy, storage mechanisms, and the complete management lifecycle from both biological and artificial perspectives. Subsequently, we review the mainstream benchmarks for evaluating agent memory. Additionally, we explore memory security from dual perspectives of attack and defense. Finally, we envision future research directions, with a focus on multimodal memory systems and skill acquisition.", "published": "2025-12-29", "updated": "2025-12-29", "authors": [ "Jiafeng Liang", "Hao Li", "Chang Li", "Jiaqi Zhou", "Shixin Jiang", "Zekun Wang", "Changkai Ji", "Zhihao Zhu", "Runxuan Liu", "Tao Ren", "Jinlan Fu", "See-Kiong Ng", "Xia Liang", "Ming Liu", "Bing Qin" ], "categories": [ "cs.CL", "cs.AI", "cs.CV" ], "url": "https://arxiv.org/abs/2512.23343", "pdf_url": "https://arxiv.org/pdf/2512.23343", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2601.01885", "title": "Agentic Memory: Learning Unified Long-Term and Short-Term Memory Management for Large Language Model Agents", "abstract": "Large language model (LLM) agents face fundamental limitations in long-horizon reasoning due to finite context windows, making effective memory management critical. Existing methods typically handle long-term memory (LTM) and short-term memory (STM) as separate components, relying on heuristics or auxiliary controllers, which limits adaptability and end-to-end optimization. In this paper, we propose Agentic Memory (AgeMem), a unified framework that integrates LTM and STM management directly into the agent's policy. AgeMem exposes memory operations as tool-based actions, enabling the LLM agent to autonomously decide what and when to store, retrieve, update, summarize, or discard information. To train such unified behaviors, we propose a three-stage progressive reinforcement learning strategy and design a step-wise GRPO to address sparse and discontinuous rewards induced by memory operations. Experiments on five long-horizon benchmarks demonstrate that AgeMem consistently outperforms strong memory-augmented baselines across multiple LLM backbones, achieving improved task performance, higher-quality long-term memory, and more efficient context usage.", "published": "2026-01-05", "updated": "2026-04-30", "authors": [ "Yi Yu", "Liuyi Yao", "Yuexiang Xie", "Qingquan Tan", "Jiaqi Feng", "Yaliang Li", "Libing Wu" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2601.01885", "pdf_url": "https://arxiv.org/pdf/2601.01885", "matched_queries": [ "agent-memory", "long-term-memory" ] }, { "arxiv_id": "2601.02577", "title": "Orchestral AI: A Framework for Agent Orchestration", "abstract": "The rapid proliferation of LLM agent frameworks has forced developers to choose between vendor lock-in through provider-specific SDKs and complex multi-package ecosystems that obscure control flow and hinder reproducibility. Integrating tool calling across multiple LLM providers remains a core engineering challenge due to fragmented APIs, incompatible message formats, and inconsistent streaming and tool-calling behavior, making it difficult to build portable, reliable agent systems. We introduce Orchestral, a lightweight Python framework that provides a unified, type-safe interface for building LLM agents across major providers while preserving the simplicity required for scientific computing and production deployment. Orchestral defines a single universal representation for messages, tools, and LLM usage that operates seamlessly across providers, eliminating manual format translation and reducing framework-induced complexity. Automatic tool schema generation from Python type hints removes the need for handwritten descriptors while maintaining type safety across provider boundaries. A synchronous execution model with streaming support enables deterministic behavior, straightforward debugging, and real-time interaction without introducing server dependencies. The framework's modular architecture cleanly separates provider integration, tool execution, conversation orchestration, and user-facing interfaces, enabling extensibility without architectural entanglement. Orchestral supports advanced agent capabilities found in larger frameworks, including rich tool calling, context compaction, workspace sandboxing, user approval workflows, sub-agents, memory management, and MCP integration.", "published": "2026-01-05", "updated": "2026-01-05", "authors": [ "Alexander Roman", "Jacob Roman" ], "categories": [ "cs.AI", "astro-ph.IM", "hep-ph" ], "url": "https://arxiv.org/abs/2601.02577", "pdf_url": "https://arxiv.org/pdf/2601.02577", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2601.02732", "title": "Agentic Memory Enhanced Recursive Reasoning for Root Cause Localization in Microservices", "abstract": "As contemporary microservice systems become increasingly popular and complex-often comprising hundreds or even thousands of fine-grained, interdependent subsystems-they are experiencing more frequent failures. Ensuring system reliability thus demands accurate root cause localization. While many traditional graph-based and deep learning approaches have been explored for this task, they often rely heavily on pre-defined schemas that struggle to adapt to evolving operational contexts. Consequently, a number of LLM-based methods have recently been proposed. However, these methods still face two major limitations: shallow, symptom-centric reasoning that undermines accuracy, and a lack of cross-alert reuse that leads to redundant reasoning and high latency. In this paper, we conduct a comprehensive study of how Site Reliability Engineers (SREs) localize the root causes of failures, drawing insights from professionals across multiple organizations. Our investigation reveals that expert root cause analysis exhibits three key characteristics: recursiveness, multi-dimensional expansion, and cross-modal reasoning. Motivated by these findings, we introduce AMER-RCL, an agentic memory enhanced recursive reasoning framework for root cause localization in microservices. AMER-RCL employs the Recursive Reasoning RCL engine, a multi-agent framework that performs recursive reasoning on each alert to progressively refine candidate causes, while Agentic Memory incrementally accumulates and reuses reasoning from prior alerts within a time window to reduce redundant exploration and lower inference latency. Experimental results demonstrate that AMER-RCL consistently outperforms state-of-the-art methods in both localization accuracy and inference efficiency.", "published": "2026-01-06", "updated": "2026-01-06", "authors": [ "Lingzhe Zhang", "Tong Jia", "Yunpeng Zhai", "Leyi Pan", "Chiming Duan", "Minghua He", "Mengxi Jia", "Ying Li" ], "categories": [ "cs.SE", "cs.AI" ], "url": "https://arxiv.org/abs/2601.02732", "pdf_url": "https://arxiv.org/pdf/2601.02732", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2601.02744", "title": "SYNAPSE: Empowering LLM Agents with Episodic-Semantic Memory via Spreading Activation", "abstract": "While Large Language Models (LLMs) excel at generalized reasoning, standard retrieval-augmented approaches fail to address the disconnected nature of long-term agentic memory. To bridge this gap, we introduce Synapse (Synergistic Associative Processing Semantic Encoding), a unified memory architecture that transcends static vector similarity. Drawing from cognitive science, Synapse models memory as a dynamic graph where relevance emerges from spreading activation rather than pre-computed links. By integrating lateral inhibition and temporal decay, the system dynamically highlights relevant sub-graphs while filtering interference. We implement a Triple Hybrid Retrieval strategy that fuses geometric embeddings with activation-based graph traversal. Comprehensive evaluations on the LoCoMo benchmark show that Synapse significantly outperforms state-of-the-art methods in complex temporal and multi-hop reasoning tasks, offering a robust solution to the \"Contextual Tunneling\" problem. Our code and data will be made publicly available upon acceptance.", "published": "2026-01-06", "updated": "2026-02-16", "authors": [ "Hanqi Jiang", "Junhao Chen", "Yi Pan", "Ling Chen", "Weihang You", "Yifan Zhou", "Ruidong Zhang", "Andrea Sikora", "Lin Zhao", "Yohannes Abate", "Tianming Liu" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2601.02744", "pdf_url": "https://arxiv.org/pdf/2601.02744", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2601.03236", "title": "MAGMA: A Multi-Graph based Agentic Memory Architecture for AI Agents", "abstract": "Memory-Augmented Generation (MAG) extends Large Language Models with external memory to support long-context reasoning, but existing approaches largely rely on semantic similarity over monolithic memory stores, entangling temporal, causal, and entity information. This design limits interpretability and alignment between query intent and retrieved evidence, leading to suboptimal reasoning accuracy. In this paper, we propose MAGMA, a multi-graph agentic memory architecture that represents each memory item across orthogonal semantic, temporal, causal, and entity graphs. MAGMA formulates retrieval as policy-guided traversal over these relational views, enabling query-adaptive selection and structured context construction. By decoupling memory representation from retrieval logic, MAGMA provides transparent reasoning paths and fine-grained control over retrieval. Experiments on LoCoMo and LongMemEval demonstrate that MAGMA consistently outperforms state-of-the-art agentic memory systems in long-horizon reasoning tasks.", "published": "2026-01-06", "updated": "2026-04-16", "authors": [ "Dongming Jiang", "Yi Li", "Guanpeng Li", "Bingzhe Li" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2601.03236", "pdf_url": "https://arxiv.org/pdf/2601.03236", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2601.03543", "title": "EvolMem: A Cognitive-Driven Benchmark for Multi-Session Dialogue Memory", "abstract": "Despite recent advances in understanding and leveraging long-range conversational memory, existing benchmarks still lack systematic evaluation of large language models(LLMs) across diverse memory dimensions, particularly in multi-session settings. In this work, we propose EvolMem, a new benchmark for assessing multi-session memory capabilities of LLMs and agent systems. EvolMem is grounded in cognitive psychology and encompasses both declarative and non-declarative memory, further decomposed into multiple fine-grained abilities. To construct the benchmark, we introduce a hybrid data synthesis framework that consists of topic-initiated generation and narrative-inspired transformations. This framework enables scalable generation of multi-session conversations with controllable complexity, accompanied by sample-specific evaluation guidelines. Extensive evaluation reveals that no LLM consistently outperforms others across all memory dimensions. Moreover, agent memory mechanisms do not necessarily enhance LLMs' capabilities and often exhibit notable efficiency limitations. Data and code will be released at https://github.com/shenye7436/EvolMem.", "published": "2026-01-07", "updated": "2026-01-07", "authors": [ "Ye Shen", "Dun Pei", "Yiqiu Guo", "Junying Wang", "Yijin Guo", "Zicheng Zhang", "Qi Jia", "Jun Zhou", "Guangtao Zhai" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2601.03543", "pdf_url": "https://arxiv.org/pdf/2601.03543", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2601.03785", "title": "Membox: Weaving Topic Continuity into Long-Range Memory for LLM Agents", "abstract": "Long-term human-agent dialogues are organized by topic continuity: adjacent turns often develop the same goal, plan, problem, or event, while related activities may recur across distant sessions. Yet many LLM agent memory systems first decompose histories into isolated turns or fixed-size chunks, then compensate through enrichment, consolidation, or retrieval mechanisms still tied to semantic proximity or fragment-level records. This weakens temporal and causal organization and biases memory access toward semantic proximity rather than task- or topic-level continuity. We introduce \\emph{Membox}, a hierarchical memory architecture that instantiates topic continuity as an explicit organization layer for agent memory. Its \\textbf{Topic Loom} incrementally organizes dialogue streams into boxes whose internal turns follow the same local topic, while its \\textbf{Trace Weaver} links extracted events across boxes into macro-topic traces that recover recurring activities, goals, and factual developments across distant sessions. On LoCoMo, Topic-Loom-only retrieval improves over the best Mem0/A-MEM retrieval-depth setting by 13.00 F1 points (53.95 vs. 40.95), and trace-expanded retrieval further raises F1 to 55.28; with GPT-4o, trace-expanded retrieval reaches 59.71 F1. Additional DialSim results show the same gain from adding cross-box traces in multi-party dialogue. These results show that local topic-continuity organization and macro-topic trace expansion improve long-range memory beyond semantic retrieval over fragmented records.", "published": "2026-01-07", "updated": "2026-06-24", "authors": [ "Dehao Tao", "Guoliang Ma", "Yongfeng Huang", "Minghu Jiang" ], "categories": [ "cs.CL", "cs.AI" ], "url": "https://arxiv.org/abs/2601.03785", "pdf_url": "https://arxiv.org/pdf/2601.03785", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2601.04170", "title": "Agent Drift: Quantifying Behavioral Degradation in Multi-Agent LLM Systems Over Extended Interactions", "abstract": "Multi-agent Large Language Model (LLM) systems have emerged as powerful architectures for complex task decomposition and collaborative problem-solving. However, their long-term behavioral stability remains largely unexamined. This study introduces the concept of agent drift, defined as the progressive degradation of agent behavior, decision quality, and inter-agent coherence over extended interaction sequences. We present a comprehensive theoretical framework for understanding drift phenomena, proposing three distinct manifestations: semantic drift (progressive deviation from original intent), coordination drift (breakdown in multi-agent consensus mechanisms), and behavioral drift (emergence of unintended strategies). We introduce the Agent Stability Index (ASI), a novel composite metric framework for quantifying drift across twelve dimensions, including response consistency, tool usage patterns, reasoning pathway stability, and inter-agent agreement rates. Through simulation-based analysis and theoretical modeling, we demonstrate how unchecked agent drift can lead to substantial reductions in task completion accuracy and increased human intervention requirements. We propose three mitigation strategies: episodic memory consolidation, drift-aware routing protocols, and adaptive behavioral anchoring. Theoretical analysis suggests these approaches can significantly reduce drift-related errors while maintaining system throughput. This work establishes a foundational methodology for monitoring, measuring, and mitigating agent drift in production agentic AI systems, with direct implications for enterprise deployment reliability and AI safety research.", "published": "2026-01-07", "updated": "2026-01-07", "authors": [ "Abhishek Rath" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2601.04170", "pdf_url": "https://arxiv.org/pdf/2601.04170", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2601.05504", "title": "Memory Poisoning Attack and Defense on Memory Based LLM-Agents", "abstract": "Large language model agents equipped with persistent memory are vulnerable to memory poisoning attacks, where adversaries inject malicious instructions through query only interactions that corrupt the agents long term memory and influence future responses. Recent work demonstrated that the MINJA (Memory Injection Attack) achieves over 95 % injection success rate and 70 % attack success rate under idealized conditions. However, the robustness of these attacks in realistic deployments and effective defensive mechanisms remain understudied. This work addresses these gaps through systematic empirical evaluation of memory poisoning attacks and defenses in Electronic Health Record (EHR) agents. We investigate attack robustness by varying three critical dimensions: initial memory state, number of indication prompts, and retrieval parameters. Our experiments on GPT-4o-mini, Gemini-2.0-Flash and Llama-3.1-8B-Instruct models using MIMIC-III clinical data reveal that realistic conditions with pre-existing legitimate memories dramatically reduce attack effectiveness. We then propose and evaluate two novel defense mechanisms: (1) Input/Output Moderation using composite trust scoring across multiple orthogonal signals, and (2) Memory Sanitization with trust-aware retrieval employing temporal decay and pattern-based filtering. Our defense evaluation reveals that effective memory sanitization requires careful trust threshold calibration to prevent both overly conservative rejection (blocking all entries) and insufficient filtering (missing subtle attacks), establishing important baselines for future adaptive defense mechanisms. These findings provide crucial insights for securing memory-augmented LLM agents in production environments.", "published": "2026-01-09", "updated": "2026-01-12", "authors": [ "Balachandra Devarangadi Sunil", "Isheeta Sinha", "Piyush Maheshwari", "Shantanu Todmal", "Shreyan Mallik", "Shuchi Mishra" ], "categories": [ "cs.CR", "cs.MA" ], "url": "https://arxiv.org/abs/2601.05504", "pdf_url": "https://arxiv.org/pdf/2601.05504", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2601.06282", "title": "Amory: Building Coherent Narrative-Driven Agent Memory through Agentic Reasoning", "abstract": "Long-term conversational agents face a fundamental scalability challenge as interactions extend over time: repeatedly processing entire conversation histories becomes computationally prohibitive. Current approaches attempt to solve this through memory frameworks that predominantly fragment conversations into isolated embeddings or graph representations and retrieve relevant ones in a RAG style. While computationally efficient, these methods often treat memory formation minimally and fail to capture the subtlety and coherence of human memory. We introduce Amory, a working memory framework that actively constructs structured memory representations through enhancing agentic reasoning during offline time. Amory organizes conversational fragments into episodic narratives, consolidates memories with momentum, and semanticizes peripheral facts into semantic memory. At retrieval time, the system employs coherence-driven reasoning over narrative structures. Evaluated on the LOCOMO benchmark for long-term reasoning, Amory achieves considerable improvements over previous state-of-the-art, with performance comparable to full context reasoning while reducing response time by 50%. Analysis shows that momentum-aware consolidation significantly enhances response quality, while coherence-driven retrieval provides superior memory coverage compared to embedding-based approaches.", "published": "2026-01-09", "updated": "2026-01-09", "authors": [ "Yue Zhou", "Xiaobo Guo", "Belhassen Bayar", "Srinivasan H. Sengamedu" ], "categories": [ "cs.CL", "cs.AI" ], "url": "https://arxiv.org/abs/2601.06282", "pdf_url": "https://arxiv.org/pdf/2601.06282", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2601.06377", "title": "HiMem: Hierarchical Long-Term Memory for LLM Long-Horizon Agents", "abstract": "Although long-term memory systems have made substantial progress in recent years, they still exhibit clear limitations in adaptability, scalability, and self-evolution under continuous interaction settings. Inspired by cognitive theories, we propose HiMem, a hierarchical long-term memory framework for long-horizon dialogues, designed to support memory construction, retrieval, and dynamic updating during sustained interactions. HiMem constructs cognitively consistent Episode Memory via a Topic-Aware Event--Surprise Dual-Channel Segmentation strategy, and builds Note Memory that captures stable knowledge through a multi-stage information extraction pipeline. These two memory types are semantically linked to form a hierarchical structure that bridges concrete interaction events and abstract knowledge, enabling efficient retrieval without sacrificing information fidelity. HiMem supports both hybrid and best-effort retrieval strategies to balance accuracy and efficiency, and incorporates conflict-aware Memory Reconsolidation to revise and supplement stored knowledge based on retrieval feedback. This design enables continual memory self-evolution over long-term use. Experimental results on long-horizon dialogue benchmarks demonstrate that HiMem consistently outperforms representative baselines in accuracy, consistency, and long-term reasoning, while maintaining favorable efficiency. Overall, HiMem provides a principled and scalable design paradigm for building adaptive and self-evolving LLM-based conversational agents. The code is available at https://github.com/jojopdq/HiMem.", "published": "2026-01-10", "updated": "2026-01-10", "authors": [ "Ningning Zhang", "Xingxing Yang", "Zhizhong Tan", "Weiping Deng", "Wenyong Wang" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2601.06377", "pdf_url": "https://arxiv.org/pdf/2601.06377", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2601.06411", "title": "Structured Episodic Event Memory", "abstract": "Current approaches to memory in Large Language Models (LLMs) predominantly rely on static Retrieval-Augmented Generation (RAG), which often results in scattered retrieval and fails to capture the structural dependencies required for complex reasoning. For autonomous agents, these passive and flat architectures lack the cognitive organization necessary to model the dynamic and associative nature of long-term interaction. To address this, we propose Structured Episodic Event Memory (SEEM), a hierarchical framework that synergizes a graph memory layer for relational facts with a dynamic episodic memory layer for narrative progression. Grounded in cognitive frame theory, SEEM transforms interaction streams into structured Episodic Event Frames (EEFs) anchored by precise provenance pointers. Furthermore, we introduce an agentic associative fusion and Reverse Provenance Expansion (RPE) mechanism to reconstruct coherent narrative contexts from fragmented evidence. Experimental results on the LoCoMo and LongMemEval benchmarks demonstrate that SEEM significantly outperforms baselines, enabling agents to maintain superior narrative coherence and logical consistency.", "published": "2026-01-10", "updated": "2026-02-10", "authors": [ "Zhengxuan Lu", "Dongfang Li", "Yukun Shi", "Beilun Wang", "Longyue Wang", "Baotian Hu" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2601.06411", "pdf_url": "https://arxiv.org/pdf/2601.06411", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2601.10744", "title": "Explore with Long-term Memory: A Benchmark and Multimodal LLM-based Reinforcement Learning Framework for Embodied Exploration", "abstract": "An ideal embodied agent should possess lifelong learning capabilities to handle long-horizon and complex tasks, enabling continuous operation in general environments. This not only requires the agent to accurately accomplish given tasks but also to leverage long-term episodic memory to optimize decision-making. However, existing mainstream one-shot embodied tasks primarily focus on task completion results, neglecting the crucial process of exploration and memory utilization. To address this, we propose Long-term Memory Embodied Exploration (LMEE), which aims to unify the agent's exploratory cognition and decision-making behaviors to promote lifelong learning. We further construct a corresponding dataset and benchmark, LMEE-Bench, incorporating multi-goal navigation and memory-based question answering to comprehensively evaluate both the process and outcome of embodied exploration. To enhance the agent's memory recall and proactive exploration capabilities, we propose MemoryExplorer, a novel method that fine-tunes a multimodal large language model through reinforcement learning to encourage active memory querying. By incorporating a multi-task reward function that includes action prediction, frontier selection, and question answering, our model achieves proactive exploration. Extensive experiments against state-of-the-art embodied exploration models demonstrate that our approach achieves significant advantages in long-horizon embodied tasks. Our dataset and code will be released at https://wangsen99.github.io/papers/lmee/", "published": "2026-01-11", "updated": "2026-03-22", "authors": [ "Sen Wang", "Bangwei Liu", "Zhenkun Gao", "Lizhuang Ma", "Xuhong Wang", "Yuan Xie", "Xin Tan" ], "categories": [ "cs.AI", "cs.CV" ], "url": "https://arxiv.org/abs/2601.10744", "pdf_url": "https://arxiv.org/pdf/2601.10744", "matched_queries": [ "agent-memory", "episodic-memory" ] }, { "arxiv_id": "2601.07470", "title": "Learning How to Remember: A Meta-Cognitive Management Method for Structured and Transferable Agent Memory", "abstract": "Large language model (LLM) agents increasingly rely on accumulated memory to solve long-horizon decision-making tasks. However, most existing approaches store memory in fixed representations and reuse it at a single or implicit level of abstraction, which limits generalization and often leads to negative transfer when distribution shift. This paper proposes the Meta-Cognitive Memory Abstraction method (MCMA), which treats memory abstraction as a learnable cognitive skill rather than a fixed design choice. MCMA decouples task execution from memory management by combining a frozen task model with a learned memory copilot. The memory copilot is trained using direct preference optimization, it determines how memories should be structured, abstracted, and reused. Memories are further organized into a hierarchy of abstraction levels, enabling selective reuse based on task similarity. When no memory is transferable, MCMA transfers the ability to abstract and manage memory by transferring the memory copilot. Experiments on ALFWorld, ScienceWorld, and BabyAI demonstrate substantial improvements in performance, out-of-distribution generalization, and cross-task transfer over several baselines.", "published": "2026-01-12", "updated": "2026-01-12", "authors": [ "Sirui Liang", "Pengfei Cao", "Jian Zhao", "Wenhao Teng", "Xiangwen Liao", "Jun Zhao", "Kang Liu" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2601.07470", "pdf_url": "https://arxiv.org/pdf/2601.07470", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2601.07779", "title": "OS-Symphony: A Holistic Framework for Robust and Generalist Computer-Using Agent", "abstract": "While Vision-Language Models (VLMs) have significantly advanced Computer-Using Agents (CUAs), current frameworks struggle with robustness in long-horizon workflows and generalization in novel domains. These limitations stem from a lack of granular control over historical visual context curation and the absence of visual-aware tutorial retrieval. To bridge these gaps, we introduce OS-Symphony, a holistic framework that comprises an Orchestrator coordinating two key innovations for robust automation: (1) a Reflection-Memory Agent that utilizes milestone-driven long-term memory to enable trajectory-level self-correction, effectively mitigating visual context loss in long-horizon tasks; (2) Versatile Tool Agents featuring a Multimodal Searcher that adopts a SeeAct paradigm to navigate a browser-based sandbox to synthesize live, visually aligned tutorials, thereby resolving fidelity issues in unseen scenarios. Experimental results demonstrate that OS-Symphony delivers substantial performance gains across varying model scales, establishing new state-of-the-art results on three online benchmarks, notably achieving 65.84% on OSWorld.", "published": "2026-01-12", "updated": "2026-01-12", "authors": [ "Bowen Yang", "Kaiming Jin", "Zhenyu Wu", "Zhaoyang Liu", "Qiushi Sun", "Zehao Li", "JingJing Xie", "Zhoumianze Liu", "Fangzhi Xu", "Kanzhi Cheng", "Qingyun Li", "Yian Wang", "Yu Qiao", "Zun Wang", "Zichen Ding" ], "categories": [ "cs.MA", "cs.AI", "cs.CL", "cs.CV", "cs.HC" ], "url": "https://arxiv.org/abs/2601.07779", "pdf_url": "https://arxiv.org/pdf/2601.07779", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2601.08160", "title": "SwiftMem: Fast Agentic Memory via Query-aware Indexing", "abstract": "Agentic memory systems have become critical for enabling LLM agents to maintain long-term context and retrieve relevant information efficiently. However, existing memory frameworks suffer from a fundamental limitation: they perform exhaustive retrieval across the entire storage layer regardless of query characteristics. This brute-force approach creates severe latency bottlenecks as memory grows, hindering real-time agent interactions. We propose SwiftMem, a query-aware agentic memory system that achieves sub-linear retrieval through specialized indexing over temporal and semantic dimensions. Our temporal index enables logarithmic-time range queries for time-sensitive retrieval, while the semantic DAG-Tag index maps queries to relevant topics through hierarchical tag structures. To address memory fragmentation during growth, we introduce an embedding-tag co-consolidation mechanism that reorganizes storage based on semantic clusters to improve cache locality. Experiments on LoCoMo and LongMemEval benchmarks demonstrate that SwiftMem achieves 47$\\times$ faster search compared to state-of-the-art baselines while maintaining competitive accuracy, enabling practical deployment of memory-augmented LLM agents.", "published": "2026-01-13", "updated": "2026-01-13", "authors": [ "Anxin Tian", "Yiming Li", "Xing Li", "Hui-Ling Zhen", "Lei Chen", "Xianzhi Yu", "Zhenhua Dong", "Mingxuan Yuan" ], "categories": [ "cs.CL", "cs.AI" ], "url": "https://arxiv.org/abs/2601.08160", "pdf_url": "https://arxiv.org/pdf/2601.08160", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2601.08323", "title": "AtomMem : Learnable Dynamic Agentic Memory with Atomic Memory Operation", "abstract": "Equipping agents with memory is essential for solving real-world long-horizon problems. However, most existing agent memory mechanisms rely on static and hand-crafted workflows. This limits the performance and generalization ability of these memory designs, which highlights the need for a more flexible, learning-based memory framework. In this paper, we propose AtomMem, which reframes memory management as a dynamic decision-making problem. We deconstruct high-level memory processes into fundamental atomic CRUD (Create, Read, Update, Delete) operations, transforming the memory workflow into a learnable decision process. By combining supervised fine-tuning with reinforcement learning, AtomMem learns an autonomous, task-aligned policy to orchestrate memory behaviors tailored to specific task demands. Experimental results across 3 long-context benchmarks demonstrate that the trained AtomMem-8B consistently outperforms prior static-workflow memory methods. Further analysis of training dynamics shows that our learning-based formulation enables the agent to discover structured, task-aligned memory management strategies, highlighting a key advantage over predefined routines.", "published": "2026-01-13", "updated": "2026-03-27", "authors": [ "Yupeng Huo", "Yaxi Lu", "Zhong Zhang", "Haotian Chen", "Yankai Lin" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2601.08323", "pdf_url": "https://arxiv.org/pdf/2601.08323", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2601.09113", "title": "The AI Hippocampus: How Far are We From Human Memory?", "abstract": "Memory plays a foundational role in augmenting the reasoning, adaptability, and contextual fidelity of modern Large Language Models and Multi-Modal LLMs. As these models transition from static predictors to interactive systems capable of continual learning and personalized inference, the incorporation of memory mechanisms has emerged as a central theme in their architectural and functional evolution. This survey presents a comprehensive and structured synthesis of memory in LLMs and MLLMs, organizing the literature into a cohesive taxonomy comprising implicit, explicit, and agentic memory paradigms. Specifically, the survey delineates three primary memory frameworks. Implicit memory refers to the knowledge embedded within the internal parameters of pre-trained transformers, encompassing their capacity for memorization, associative retrieval, and contextual reasoning. Recent work has explored methods to interpret, manipulate, and reconfigure this latent memory. Explicit memory involves external storage and retrieval components designed to augment model outputs with dynamic, queryable knowledge representations, such as textual corpora, dense vectors, and graph-based structures, thereby enabling scalable and updatable interaction with information sources. Agentic memory introduces persistent, temporally extended memory structures within autonomous agents, facilitating long-term planning, self-consistency, and collaborative behavior in multi-agent systems, with relevance to embodied and interactive AI. Extending beyond text, the survey examines the integration of memory within multi-modal settings, where coherence across vision, language, audio, and action modalities is essential. Key architectural advances, benchmark tasks, and open challenges are discussed, including issues related to memory capacity, alignment, factual consistency, and cross-system interoperability.", "published": "2026-01-14", "updated": "2026-01-14", "authors": [ "Zixia Jia", "Jiaqi Li", "Yipeng Kang", "Yuxuan Wang", "Tong Wu", "Quansen Wang", "Xiaobo Wang", "Shuyi Zhang", "Junzhe Shen", "Qing Li", "Siyuan Qi", "Yitao Liang", "Di He", "Zilong Zheng", "Song-Chun Zhu" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2601.09113", "pdf_url": "https://arxiv.org/pdf/2601.09113", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2601.09636", "title": "PersonalAlign: Hierarchical Implicit Intent Alignment for Personalized GUI Agent with Long-Term User-Centric Records", "abstract": "While GUI agents have shown strong performance under explicit and completion instructions, real-world deployment requires aligning with users' more complex implicit intents. In this work, we highlight Hierarchical Implicit Intent Alignment for Personalized GUI Agent (PersonalAlign), a new agent task that requires agents to leverage long-term user records as persistent context to resolve omitted preferences in vague instructions and anticipate latent routines by user state for proactive assistance. To facilitate this study, we introduce AndroidIntent, a benchmark designed to evaluate agents' ability in resolving vague instructions and providing proactive suggestions through reasoning over long-term user records. We annotated 775 user-specific preferences and 215 routines from 20k long-term records across different users for evaluation. Furthermore, we introduce Hierarchical Intent Memory Agent (HIM-Agent), which maintains a continuously updating personal memory and hierarchically organizes user preferences and routines for personalization. Finally, we evaluate a range of GUI agents on AndroidIntent, including GPT-5, Qwen3-VL, and UI-TARS, further results show that HIM-Agent significantly improves both execution and proactive performance by 15.7% and 7.3%.", "published": "2026-01-14", "updated": "2026-05-13", "authors": [ "Yibo Lyu", "Gongwei Chen", "Rui Shao", "Weili Guan", "Liqiang Nie" ], "categories": [ "cs.AI", "cs.CV", "cs.HC", "cs.LG" ], "url": "https://arxiv.org/abs/2601.09636", "pdf_url": "https://arxiv.org/pdf/2601.09636", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2602.06051", "title": "CAST: Character-and-Scene Episodic Memory for Agents", "abstract": "Episodic memory is a central component of human memory, which refers to the ability to recall coherent events grounded in who, when, and where. However, most agent memory systems only emphasize semantic recall and treat experience as structures such as key-value, vector, or graph, which makes them struggle to represent and retrieve coherent events. To address this challenge, we propose a Character-and-Scene based memory architecture(CAST) inspired by dramatic theory. Specifically, CAST constructs 3D scenes (time/place/topic) and organizes them into character profiles that summarize the events of a character to represent episodic memory. Moreover, CAST complements this episodic memory with a graph-based semantic memory, which yields a robust dual memory design. Experiments demonstrate that CAST has averagely improved 8.11% F1 and 10.21% J(LLM-as-a-Judge) than baselines on various datasets, especially on open and time-sensitive conversational questions.", "published": "2026-01-14", "updated": "2026-02-18", "authors": [ "Kexin Ma", "Bojun Li", "Yuhua Tang", "Liting Sun", "Ruochun Jin" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2602.06051", "pdf_url": "https://arxiv.org/pdf/2602.06051", "matched_queries": [ "agent-memory", "episodic-memory" ] }, { "arxiv_id": "2602.06052", "title": "Rethinking Memory Mechanisms of Foundation Agents in the Second Half: A Survey", "abstract": "The research of artificial intelligence is undergoing a paradigm shift from prioritizing model innovations over benchmark scores towards emphasizing problem definition and rigorous real-world evaluation. As the field enters the \"second half,\" the central challenge becomes real utility in long-horizon, dynamic, and user-dependent environments, where agents face context explosion and must continuously accumulate, manage, and selectively reuse large volumes of information across extended interactions. Memory, with hundreds of papers released this year, therefore emerges as the critical solution to fill the utility gap. In this survey, we provide a unified view of foundation agent memory along three dimensions: memory substrate (internal and external), cognitive mechanism (episodic, semantic, sensory, working, and procedural), and memory subject (agent- and user-centric). We then analyze how memory is instantiated and operated under different agent topologies and highlight learning policies over memory operations. Finally, we review evaluation benchmarks and metrics for assessing memory utility, and outline various open challenges and future directions.", "published": "2026-01-14", "updated": "2026-02-10", "authors": [ "Wei-Chieh Huang", "Weizhi Zhang", "Yueqing Liang", "Yuanchen Bei", "Yankai Chen", "Tao Feng", "Xinyu Pan", "Zhen Tan", "Yu Wang", "Tianxin Wei", "Shanglin Wu", "Ruiyao Xu", "Liangwei Yang", "Rui Yang", "Wooseong Yang", "Chin-Yuan Yeh", "Hanrong Zhang", "Haozhen Zhang", "Siqi Zhu", "Henry Peng Zou", "Wanjia Zhao", "Song Wang", "Wujiang Xu", "Zixuan Ke", "Zheng Hui", "Dawei Li", "Yaozu Wu", "Langzhou He", "Chen Wang", "Xiongxiao Xu", "Baixiang Huang", "Juntao Tan", "Shelby Heinecke", "Huan Wang", "Caiming Xiong", "Ahmed A. Metwally", "Jun Yan", "Chen-Yu Lee", "Hanqing Zeng", "Yinglong Xia", "Xiaokai Wei", "Ali Payani", "Yu Wang", "Haitong Ma", "Wenya Wang", "Chenguang Wang", "Yu Zhang", "Xin Wang", "Yongfeng Zhang", "Jiaxuan You", "Hanghang Tong", "Xiao Luo", "Xue Liu", "Yizhou Sun", "Wei Wang", "Julian McAuley", "James Zou", "Jiawei Han", "Philip S. Yu", "Kai Shu" ], "categories": [ "cs.CL", "cs.AI" ], "url": "https://arxiv.org/abs/2602.06052", "pdf_url": "https://arxiv.org/pdf/2602.06052", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2601.10702", "title": "Grounding Agent Memory in Contextual Intent", "abstract": "Deploying large language models in long-horizon, goal-oriented interactions remains challenging because similar entities and facts recur under different latent goals and constraints, causing memory systems to retrieve context-mismatched evidence. We propose STITCH (Structured Intent Tracking in Contextual History), an agentic memory system that indexes each trajectory step with a structured retrieval cue, contextual intent, and retrieves history by matching the current step's intent. Contextual intent provides compact signals that disambiguate repeated mentions and reduce interference: (1) the current latent goal defining a thematic segment, (2) the action type, and (3) the salient entity types anchoring which attributes matter. During inference, STITCH filters and prioritizes memory snippets by intent compatibility, suppressing semantically similar but context-incompatible history. For evaluation, we introduce CAME-Bench, a benchmark for context-aware retrieval in realistic, dynamic, goal-oriented trajectories. Across CAME-Bench and LongMemEval, STITCH achieves state-of-the-art performance, outperforming the strongest baseline by 35.6%, with the largest gains as trajectory length increases. Our analysis shows that intent indexing substantially reduces retrieval noise, supporting intent-aware memory for robust long-horizon reasoning.", "published": "2026-01-15", "updated": "2026-04-30", "authors": [ "Ruozhen Yang", "Yucheng Jiang", "Yueqi Jiang", "Priyanka Kargupta", "Yunyi Zhang", "Jiawei Han" ], "categories": [ "cs.CL", "cs.AI", "cs.IR" ], "url": "https://arxiv.org/abs/2601.10702", "pdf_url": "https://arxiv.org/pdf/2601.10702", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2601.11854", "title": "ATOD: An Evaluation Framework and Benchmark for Agentic Task-Oriented Dialogue Systems", "abstract": "Recent advances in task-oriented dialogue (TOD) systems, driven by large language models (LLMs) with extensive API and tool integration, have enabled conversational agents to coordinate interleaved goals, maintain long-horizon context, and act proactively through asynchronous execution. These capabilities extend beyond traditional TOD systems, yet existing benchmarks lack systematic support for evaluating such agentic behaviors. To address this gap, we introduce ATOD, a benchmark and synthetic dialogue generation pipeline that produces richly annotated conversations requiring long-term reasoning. ATOD captures key characteristics of advanced TOD, including multi-goal coordination, dependency management, memory, adaptability, and proactivity. Building on ATOD, we propose ATOD-Eval, a holistic evaluation framework that translates these dimensions into fine-grained metrics and supports reproducible offline and online evaluation. We further present a strong agentic memory-based evaluator for benchmarking on ATOD. Experiments show that ATOD-Eval enables comprehensive assessment across task completion, agentic capability, and response quality, and that the proposed evaluator offers a better accuracy-efficiency tradeoff compared to existing memory- and LLM-based approaches under this evaluation setting.", "published": "2026-01-17", "updated": "2026-01-30", "authors": [ "Yifei Zhang", "Hooshang Nayyeri", "Rinat Khaziev", "Emine Yilmaz", "Gokhan Tur", "Dilek Hakkani-Tür", "Hari Thadakamalla" ], "categories": [ "cs.CL", "cs.AI", "cs.MA" ], "url": "https://arxiv.org/abs/2601.11854", "pdf_url": "https://arxiv.org/pdf/2601.11854", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2601.12771", "title": "Who Does This Name Remind You of ? Nationality Prediction via Large Language Model Associative Memory", "abstract": "Large language models (LLMs) possess extensive world knowledge, yet methods for effectively eliciting this knowledge remain underexplored. Nationality and region prediction tasks require understanding of not only linguistic features but also cultural and historical background, making LLM world knowledge particularly valuable. However, conventional LLM prompting methods rely on direct reasoning approaches, which have limitations in applying abstract linguistic rules. We propose LLM Associative Memory Agents (LAMA), a novel framework that leverages LLM world knowledge as associative memory. Rather than directly inferring nationality from names, LAMA recalls famous individuals with the same name and aggregates their nationalities through indirect reasoning. A dual-agent architecture comprising a Person Agent and a Media Agent, specialized in different knowledge domains, recalls famous individuals in parallel, generating Top-1 predictions through voting and Top-K predictions through conditional completion. On a 99-country nationality prediction task, LAMA achieved 0.817 accuracy, substantially outperforming conventional LLM prompting methods and neural models. Our experiments reveal that LLMs exhibit higher reliability in recalling concrete examples than in abstract reasoning, that recall-based approaches are robust to low-frequency nationalities independent of data frequency distributions, and that the dual-agent architecture functions complementarily to produce synergistic effects. These results demonstrate the effectiveness of a new multi-agent system that retrieves and aggregates LLM knowledge rather than prompting reasoning.", "published": "2026-01-19", "updated": "2026-01-23", "authors": [ "Keito Inoshita" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2601.12771", "pdf_url": "https://arxiv.org/pdf/2601.12771", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2601.14192", "title": "Toward Efficient Agents: Memory, Tool learning, and Planning", "abstract": "Recent years have witnessed increasing interest in extending large language models into agentic systems. While the effectiveness of agents has continued to improve, efficiency, which is crucial for real-world deployment, has often been overlooked. This paper therefore investigates efficiency from three core components of agents: memory, tool learning, and planning, considering costs such as latency, tokens, steps, etc. Aimed at conducting comprehensive research addressing the efficiency of the agentic system itself, we review a broad range of recent approaches that differ in implementation yet frequently converge on shared high-level principles including but not limited to bounding context via compression and management, designing reinforcement learning rewards to minimize tool invocation, and employing controlled search mechanisms to enhance efficiency, which we discuss in detail. Accordingly, we characterize efficiency in two complementary ways: comparing effectiveness under a fixed cost budget, and comparing cost at a comparable level of effectiveness. This trade-off can also be viewed through the Pareto frontier between effectiveness and cost. From this perspective, we also examine efficiency oriented benchmarks by summarizing evaluation protocols for these components and consolidating commonly reported efficiency metrics from both benchmark and methodological studies. Moreover, we discuss the key challenges and future directions, with the goal of providing promising insights.", "published": "2026-01-20", "updated": "2026-07-03", "authors": [ "Xiaofang Yang", "Lijun Li", "Heng Zhou", "Tong Zhu", "Xiaoye Qu", "Yuchen Fan", "Qianshan Wei", "Rui Ye", "Li Kang", "Yiran Qin", "Daizong Liu", "Qi Li", "Ning Ding", "Siheng Chen", "Jing Shao" ], "categories": [ "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2601.14192", "pdf_url": "https://arxiv.org/pdf/2601.14192", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2601.14735", "title": "Optimizing FaaS Platforms for MCP-enabled Agentic Workflows", "abstract": "Agentic workflows that use autonomous AI Agents powered by Large Language Models (LLMs) and Model Context Protocol (MCP) servers is rapidly rising. This introduces challenges in scalable cloud deployment and state management. Traditional hosting on Virtual Machines (VMs) is resource-intensive and lacks elasticity. Functions-as-a-Service (FaaS) platforms offer modularity, autoscaling and cost efficiency but are inherently stateless. In this paper, we present the FAME, a FaaS-based architecture for orchestrating MCP-enabled agentic workflows. FAME decomposes agentic patterns such as ReAct into composable agents: Planner, Actor and Evaluator, that are each a FaaS function built using LangGraph and are orchestrated as a FaaS workflow. This enables modular composition as AWS Step Functions and avoids function timeouts seen for monolithic agentic workflows. To address context persistence across user requests in a conversation, FAME automates agent memory persistence and injection using DynamoDB. It also optimizes MCP server deployment through AWS Lambda wrappers, caches tool outputs in S3 and proposes function fusion strategies. We evaluate FAME on two representative applications, on research paper summarization and log analytics, under diverse memory and caching configurations. Results show up to 13x latency reduction, 88% fewer input tokens and 66% in cost savings, along with improved workflow completion rates. This demonstrates the viability of serverless platforms for hosting complex, multi-agent AI workflows at scale.", "published": "2026-01-21", "updated": "2026-01-27", "authors": [ "Varad Kulkarni", "Vaibhav Jha", "Nikhil Reddy", "Anand Eswaran", "Praveen Jayachandran", "Yogesh Simmhan" ], "categories": [ "cs.DC" ], "url": "https://arxiv.org/abs/2601.14735", "pdf_url": "https://arxiv.org/pdf/2601.14735", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2601.16690", "title": "EMemBench: Interactive Benchmarking of Episodic Memory for VLM Agents", "abstract": "We introduce EMemBench, a programmatic benchmark for evaluating long-term memory of agents through interactive games. Rather than using a fixed set of questions, EMemBench generates questions from each agent's own trajectory, covering both text and visual game environments. Each template computes verifiable ground truth from underlying game signals, with controlled answerability and balanced coverage over memory skills: single/multi-hop recall, induction, temporal, spatial, logical, and adversarial. We evaluate memory agents with strong LMs/VLMs as backbones, using in-context prompting as baselines. Across 15 text games and multiple visual seeds, results are far from saturated: induction and spatial reasoning are persistent bottlenecks, especially in visual setting. Persistent memory yields clear gains for open backbones on text games, but improvements are less consistent for VLM agents, suggesting that visually grounded episodic memory remains an open challenge. A human study further confirms the difficulty of EMemBench.", "published": "2026-01-23", "updated": "2026-01-23", "authors": [ "Xinze Li", "Ziyue Zhu", "Siyuan Liu", "Yubo Ma", "Yuhang Zang", "Yixin Cao", "Aixin Sun" ], "categories": [ "cs.CL", "cs.CV" ], "url": "https://arxiv.org/abs/2601.16690", "pdf_url": "https://arxiv.org/pdf/2601.16690", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2601.16872", "title": "From Atom to Community: Structured and Evolving Agent Memory for User Behavior Modeling", "abstract": "User behavior modeling lies at the heart of personalized applications like recommender systems. With LLM-based agents, user preference representation has evolved from latent embeddings to semantic memory. While existing memory mechanisms show promise in textual dialogues, modeling non-textual behaviors remains challenging, as preferences must be inferred from implicit signals like clicks without ground truth supervision. Current approaches rely on a single unstructured summary, updated through simple overwriting. However, this is suboptimal: users exhibit multi-faceted interests that get conflated, preferences evolve yet naive overwriting causes forgetting, and sparse individual interactions necessitate collaborative signals. We present STEAM (\\textit{\\textbf{ST}ructured and \\textbf{E}volving \\textbf{A}gent \\textbf{M}emory}), a novel framework that reimagines how agent memory is organized and updated. STEAM decomposes preferences into atomic memory units, each capturing a distinct interest dimension with explicit links to observed behaviors. To exploit collaborative patterns, STEAM organizes similar memories across users into communities and generates prototype memories for signal propagation. The framework further incorporates adaptive evolution mechanisms, including consolidation for refining memories and formation for capturing emerging interests. Experiments on three real-world datasets demonstrate that STEAM substantially outperforms state-of-the-art baselines in recommendation accuracy, simulation fidelity, and diversity.", "published": "2026-01-23", "updated": "2026-01-26", "authors": [ "Yuxin Liao", "Le Wu", "Min Hou", "Yu Wang", "Han Wu", "Meng Wang" ], "categories": [ "cs.IR" ], "url": "https://arxiv.org/abs/2601.16872", "pdf_url": "https://arxiv.org/pdf/2601.16872", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2601.17887", "title": "When Personalization Legitimizes Risks: Uncovering Safety Vulnerabilities in Personalized Dialogue Agents", "abstract": "Long-term memory enables large language model (LLM) agents to support personalized and sustained interactions. However, most work on personalized agents prioritizes utility and user experience, treating memory as a neutral component and largely overlooking its safety implications. In this paper, we reveal intent legitimation, a previously underexplored safety failure in personalized agents, where benign personal memories bias intent inference and cause models to legitimize inherently harmful queries. To study this phenomenon, we introduce PS-Bench, a benchmark designed to identify and quantify intent legitimation in personalized interactions. Across multiple memory-augmented agent frameworks and base LLMs, personalization increases attack success rates by 15.8\\%--243.7\\% relative to stateless baselines. We further provide mechanistic evidence for intent legitimation from internal representations space, and propose a lightweight detection-reflection method that effectively reduces safety degradation. Overall, our work provides the first systematic exploration and evaluation of intent legitimation as a safety failure mode that naturally arises from benign, real-world personalization, highlighting the importance of assessing safety under long-term personal context. Our code is available at: https://github.com/MuyuenLP/PS-Bench. WARNING: This paper may contain harmful content.", "published": "2026-01-25", "updated": "2026-05-17", "authors": [ "Jiahe Guo", "Xiangran Guo", "Yulin Hu", "Zimo Long", "Xingyu Sui", "Xuda Zhi", "Yongbo Huang", "Hao He", "Weixiang Zhao", "Yanyan Zhao", "Bing Qin" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2601.17887", "pdf_url": "https://arxiv.org/pdf/2601.17887", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2601.18642", "title": "FadeMem: Biologically-Inspired Forgetting for Efficient Agent Memory", "abstract": "Large language models deployed as autonomous agents face critical memory limitations, lacking selective forgetting mechanisms that lead to either catastrophic forgetting at context boundaries or information overload within them. While human memory naturally balances retention and forgetting through adaptive decay processes, current AI systems employ binary retention strategies that preserve everything or lose it entirely. We propose FadeMem, a biologically-inspired agent memory architecture that incorporates active forgetting mechanisms mirroring human cognitive efficiency. FadeMem implements differential decay rates across a dual-layer memory hierarchy, where retention is governed by adaptive exponential decay functions modulated by semantic relevance, access frequency, and temporal patterns. Through LLM-guided conflict resolution and intelligent memory fusion, our system consolidates related information while allowing irrelevant details to fade. Experiments on Multi-Session Chat, LoCoMo, and LTI-Bench demonstrate superior multi-hop reasoning and retrieval with 45\\% storage reduction, validating the effectiveness of biologically-inspired forgetting in agent memory systems.", "published": "2026-01-26", "updated": "2026-02-06", "authors": [ "Lei Wei", "Xiao Peng", "Xu Dong", "Niantao Xie", "Bin Wang" ], "categories": [ "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2601.18642", "pdf_url": "https://arxiv.org/pdf/2601.18642", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2601.20352", "title": "AMA: Adaptive Memory via Multi-Agent Collaboration", "abstract": "The rapid evolution of Large Language Model (LLM) agents has necessitated robust memory systems to support cohesive long-term interaction and complex reasoning. Benefiting from the strong capabilities of LLMs, recent research focus has shifted from simple context extension to the development of dedicated agentic memory systems. However, existing approaches typically rely on rigid retrieval granularity, accumulation-heavy maintenance strategies, and coarse-grained update mechanisms. These design choices create a persistent mismatch between stored information and task-specific reasoning demands, while leading to the unchecked accumulation of logical inconsistencies over time. To address these challenges, we propose Adaptive Memory via Multi-Agent Collaboration (AMA), a novel framework that leverages coordinated agents to manage memory across multiple granularities. AMA employs a hierarchical memory design that dynamically aligns retrieval granularity with task complexity. Specifically, the Constructor and Retriever jointly enable multi-granularity memory construction and adaptive query routing. The Judge verifies the relevance and consistency of retrieved content, triggering iterative retrieval when evidence is insufficient or invoking the Refresher upon detecting logical conflicts. The Refresher then enforces memory consistency by performing targeted updates or removing outdated entries. Extensive experiments on challenging long-context benchmarks show that AMA significantly outperforms state-of-the-art baselines while reducing token consumption by approximately 80% compared to full-context methods, demonstrating its effectiveness in maintaining retrieval precision and long-term memory consistency.", "published": "2026-01-28", "updated": "2026-04-15", "authors": [ "Weiquan Huang", "Zixuan Wang", "Hehai Lin", "Sudong Wang", "Bo Xu", "Qian Li", "Beier Zhu", "Linyi Yang", "Chengwei Qin" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2601.20352", "pdf_url": "https://arxiv.org/pdf/2601.20352", "matched_queries": [ "agent-memory", "long-term-memory" ] }, { "arxiv_id": "2601.20465", "title": "BMAM: Brain-inspired Multi-Agent Memory Framework", "abstract": "Language-model-based agents operating over extended interaction horizons face persistent challenges in preserving temporally grounded information and maintaining behavioral consistency across sessions, a failure mode we term soul erosion. We present BMAM (Brain-inspired Multi-Agent Memory), a general-purpose memory architecture that models agent memory as a set of functionally specialized subsystems rather than a single unstructured store. Inspired by cognitive memory systems, BMAM decomposes memory into episodic, semantic, salience-aware, and control-oriented components that operate at complementary time scales. To support long-horizon reasoning, BMAM organizes episodic memories along explicit timelines and retrieves evidence by fusing multiple complementary signals. Experiments on the LoCoMo benchmark show that BMAM achieves 78.45 percent accuracy under the standard long-horizon evaluation setting, and ablation analyses confirm that the hippocampus-inspired episodic memory subsystem plays a critical role in temporal reasoning.", "published": "2026-01-28", "updated": "2026-01-28", "authors": [ "Yang Li", "Jiaxiang Liu", "Yusong Wang", "Yujie Wu", "Mingkun Xu" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2601.20465", "pdf_url": "https://arxiv.org/pdf/2601.20465", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2601.21468", "title": "MemOCR: Layout-Aware Visual Memory for Efficient Long-Horizon Reasoning", "abstract": "Long-horizon agentic reasoning necessitates effectively compressing growing interaction histories into a limited context window. Most existing memory systems serialize history as text, where token-level cost is uniform and scales linearly with length, often spending scarce budget on low-value details. To this end, we introduce MemOCR, a multimodal memory agent that improves long-horizon reasoning under tight context budgets by allocating memory space with adaptive information density through visual layout. Concretely, MemOCR maintains a structured rich-text memory (e.g., headings, highlights) and renders it into an image that the agent consults for memory access, visually prioritizing crucial evidence while aggressively compressing auxiliary details. To ensure robustness across varying memory budgets, we train MemOCR with reinforcement learning under budget-aware objectives that expose the agent to diverse compression levels. Across long-context multi-hop and single-hop question-answering benchmarks, MemOCR outperforms strong text-based baselines and achieves more effective context utilization under extreme budgets.", "published": "2026-01-29", "updated": "2026-05-18", "authors": [ "Yaorui Shi", "Shugui Liu", "Yu Yang", "Wenyu Mao", "Yuxin Chen", "Qi GU", "Hui Su", "Xunliang Cai", "Xiang Wang", "An Zhang" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2601.21468", "pdf_url": "https://arxiv.org/pdf/2601.21468", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2601.21714", "title": "E-mem: Multi-agent based Episodic Context Reconstruction for LLM Agent Memory", "abstract": "The evolution of Large Language Model (LLM) agents towards System~2 reasoning, characterized by deliberative, high-precision problem-solving, requires maintaining rigorous logical integrity over extended horizons. However, prevalent memory preprocessing paradigms suffer from destructive de-contextualization. By compressing complex sequential dependencies into pre-defined structures (e.g., embeddings or graphs), these methods sever the contextual integrity essential for deep reasoning. To address this, we propose E-mem, a framework shifting from Memory Preprocessing to Episodic Context Reconstruction. Inspired by biological engrams, E-mem employs a heterogeneous hierarchical architecture where multiple assistant agents maintain uncompressed memory contexts, while a central master agent orchestrates global planning. Unlike passive retrieval, our mechanism empowers assistants to locally reason within activated segments, extracting context-aware evidence before aggregation. Evaluations on the LoCoMo benchmark demonstrate that E-mem achieves over 54\\% F1, surpassing the state-of-the-art GAM by 7.75\\%, while reducing token cost by over 70\\%.", "published": "2026-01-29", "updated": "2026-06-13", "authors": [ "Kaixiang Wang", "Yidan Lin", "Jiong Lou", "Zhaojiacheng Zhou", "Bunyod Suvonov", "Jie Li" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2601.21714", "pdf_url": "https://arxiv.org/pdf/2601.21714", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2601.21841", "title": "Embodied Task Planning via Graph-Informed Action Generation with Large Language Models", "abstract": "While Large Language Models (LLMs) have demonstrated strong zero-shot reasoning capabilities, their deployment as embodied agents still faces fundamental challenges in long-horizon planning. Unlike open-ended text generation, embodied agents must decompose high-level intents into actionable sub-goals while adhering to the constraints of a dynamic environment. Standard LLM planners frequently fail to maintain strategy coherence over extended horizons due to context window limitations or hallucinate state transitions that violate environment constraints. We propose GiG, a planning framework that structures embodied agents' memory using a Graph-in-Graph architecture. Our approach employs a Graph Neural Network (GNN) to encode environmental states into embeddings, organizing these embeddings into action-connected execution trace graphs within an experience memory bank. GiG enables retrieval of structurally-similar priors, allowing agents to ground current decisions in relevant past structural patterns. Furthermore, we introduce a bounded lookahead module that leverages symbolic transition logic to enhance the agent's planning capabilities through grounded action projections. We evaluate our framework on three embodied planning benchmarks-Robotouille Synchronous, Robotouille Asynchronous, and ALFWorld. Our method outperforms state-of-the-art baselines, achieving Pass@1 performance gains of up to 22% on Robotouille Synchronous, 37% on Asynchronous, and 15% on ALFWorld while maintaining comparable or lower computational cost.", "published": "2026-01-29", "updated": "2026-05-17", "authors": [ "Xiang Li", "Ning Yan", "Masood Mortazavi" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2601.21841", "pdf_url": "https://arxiv.org/pdf/2601.21841", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2601.22964", "title": "EvoClinician: A Self-Evolving Agent for Multi-Turn Medical Diagnosis via Test-Time Evolutionary Learning", "abstract": "Prevailing medical AI operates on an unrealistic ''one-shot'' model, diagnosing from a complete patient file. However, real-world diagnosis is an iterative inquiry where Clinicians sequentially ask questions and order tests to strategically gather information while managing cost and time. To address this, we first propose Med-Inquire, a new benchmark designed to evaluate an agent's ability to perform multi-turn diagnosis. Built upon a dataset of real-world clinical cases, Med-Inquire simulates the diagnostic process by hiding a complete patient file behind specialized Patient and Examination agents. They force the agent to proactively ask questions and order tests to gather information piece by piece. To tackle the challenges posed by Med-Inquire, we then introduce EvoClinician, a self-evolving agent that learns efficient diagnostic strategies at test time. Its core is a ''Diagnose-Grade-Evolve'' loop: an Actor agent attempts a diagnosis; a Process Grader agent performs credit assignment by evaluating each action for both clinical yield and resource efficiency; finally, an Evolver agent uses this feedback to update the Actor's strategy by evolving its prompt and memory. Our experiments show EvoClinician outperforms continual learning baselines and other self-evolving agents like memory agents. The code is available at https://github.com/yf-he/EvoClinician", "published": "2026-01-30", "updated": "2026-01-30", "authors": [ "Yufei He", "Juncheng Liu", "Zhiyuan Hu", "Yulin Chen", "Yue Liu", "Yuan Sui", "Yibo Li", "Nuo Chen", "Jun Hu", "Bryan Hooi", "Xinxing Xu", "Jiang Bian" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2601.22964", "pdf_url": "https://arxiv.org/pdf/2601.22964", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2601.22974", "title": "MiTa: A Hierarchical Multi-Agent Collaboration Framework with Memory-integrated and Task Allocation", "abstract": "Recent advances in large language models (LLMs) have substantially accelerated the development of embodied agents. LLM-based multi-agent systems mitigate the inefficiency of single agents in complex tasks. However, they still suffer from issues such as memory inconsistency and agent behavioral conflicts. To address these challenges, we propose MiTa, a hierarchical memory-integrated task allocative framework to enhance collaborative efficiency. MiTa organizes agents into a manager-member hierarchy, where the manager incorporates additional allocation and summary modules that enable (1) global task allocation and (2) episodic memory integration. The allocation module enables the manager to allocate tasks from a global perspective, thereby avoiding potential inter-agent conflicts. The summary module, triggered by task progress updates, performs episodic memory integration by condensing recent collaboration history into a concise summary that preserves long-horizon context. By combining task allocation with episodic memory, MiTa attains a clearer understanding of the task and facilitates globally consistent task distribution. Experimental results confirm that MiTa achieves superior efficiency and adaptability in complex multi-agent cooperation over strong baseline methods.", "published": "2026-01-30", "updated": "2026-01-30", "authors": [ "XiaoJie Zhang", "JianHan Wu", "Xiaoyang Qu", "Jianzong Wang" ], "categories": [ "cs.ET", "cs.CL" ], "url": "https://arxiv.org/abs/2601.22974", "pdf_url": "https://arxiv.org/pdf/2601.22974", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2601.23014", "title": "Mem-T: Densifying Rewards for Long-Horizon Memory Agents", "abstract": "Memory agents, which depart from predefined memory-processing pipelines by endogenously managing the processing, storage, and retrieval of memories, have garnered increasing attention for their autonomy and adaptability. However, existing training paradigms remain constrained: agents often traverse long-horizon sequences of memory operations before receiving sparse and delayed rewards, which hinders truly end-to-end optimization of memory management policies. To address this limitation, we introduce Mem-T, an autonomous memory agent that interfaces with a lightweight hierarchical memory database to perform dynamic updates and multi-turn retrieval over streaming inputs. To effectively train long-horizon memory management capabilities, we further propose MoT-GRPO, a tree-guided reinforcement learning framework that transforms sparse terminal feedback into dense, step-wise supervision via memory operation tree backpropagation and hindsight credit assignment, thereby enabling the joint optimization of memory construction and retrieval. Extensive experiments demonstrate that Mem-T is (1) high-performing, surpassing frameworks such as A-Mem and Mem0 by up to $14.92\\%$, and (2) economical, operating on a favorable accuracy-efficiency Pareto frontier and reducing inference tokens per query by $\\sim24.45\\%$ relative to GAM without sacrificing performance.", "published": "2026-01-30", "updated": "2026-03-09", "authors": [ "Yanwei Yue", "Boci Peng", "Xuanbo Fan", "Jiaxin Guo", "Qiankun Li", "Yan Zhang" ], "categories": [ "cs.LG", "cs.CL" ], "url": "https://arxiv.org/abs/2601.23014", "pdf_url": "https://arxiv.org/pdf/2601.23014", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2602.00352", "title": "DETOUR: An Interactive Benchmark for Dual-Agent Search and Reasoning", "abstract": "When recalling information in conversation, people often arrive at the recollection after multiple turns. However, existing benchmarks for evaluating agent capabilities in such tip-of-the-tongue search processes are restricted to single-turn settings. To more realistically simulate tip-of-the-tongue search, we introduce Dual-agent based Evaluation Through Obscure Under-specified Retrieval (DETOUR), a dual-agent evaluation benchmark containing 1,011 prompts. The benchmark design involves a Primary Agent, which is the subject of evaluation, tasked with identifying the recollected entity through querying a Memory Agent that is held consistent across evaluations. Our results indicate that current state-of-the-art models still struggle with our benchmark, only achieving 36% accuracy when evaluated on all modalities (text, image, audio, and video), highlighting the importance of enhancing capabilities in underspecified scenarios.", "published": "2026-01-30", "updated": "2026-01-30", "authors": [ "Li Siyan", "Darshan Deshpande", "Anand Kannappan", "Rebecca Qian" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2602.00352", "pdf_url": "https://arxiv.org/pdf/2602.00352", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2602.00364", "title": "\"Someone Hid It\": Query-Agnostic Black-Box Attacks on LLM-Based Retrieval", "abstract": "Large language models (LLMs) have been serving as effective backbones for retrieval systems, including Retrieval-Augmentation-Generation (RAG), Dense Information Retriever (IR), and Agent Memory Retrieval. Recent studies have demonstrated that such LLM-based Retrieval (LLMR) is vulnerable to adversarial attacks, which manipulates documents by token-level injections and enables adversaries to either boost or diminish these documents in retrieval tasks. However, existing attack studies mainly (1) presume a known query is given to the attacker, and (2) highly rely on access to the victim model's parameters or interactions, which are hardly accessible in real-world scenarios, leading to limited validity. To further explore the secure risks of LLMR, we propose a practical black-box attack method that generates transferable injection tokens based on zero-shot surrogate LLMs without need of victim queries or victim models knowledge. The effectiveness of our attack raises such a robustness issue that similar effects may arise from benign or unintended document edits in the real world. To achieve our attack, we first establish a theoretical framework of LLMR and empirically verify it. Under the framework, we simulate the transferable attack as a min-max problem, and propose an adversarial learning mechanism that finds optimal adversarial tokens with learnable query samples. Our attack is validated to be effective on benchmark datasets across popular LLM retrievers.", "published": "2026-01-30", "updated": "2026-05-15", "authors": [ "Jiate Li", "Defu Cao", "Li Li", "Wei Yang", "Yuehan Qin", "Chenxiao Yu", "Tiannuo Yang", "Ryan A. Rossi", "Yan Liu", "Xiyang Hu", "Yue Zhao" ], "categories": [ "cs.CR" ], "url": "https://arxiv.org/abs/2602.00364", "pdf_url": "https://arxiv.org/pdf/2602.00364", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2602.00675", "title": "Factored Reasoning with Inner Speech and Persistent Memory for Evidence-Grounded Human-Robot Interaction", "abstract": "Dialogue-based human-robot interaction requires robot cognitive assistants to maintain persistent user context, recover from underspecified requests, and ground responses in external evidence, while keeping intermediate decisions verifiable. In this paper we introduce JANUS, a cognitive architecture for assistive robots that models interaction as a partially observable Markov decision process and realizes control as a factored controller with typed interfaces. To this aim, Janus (i) decomposes the overall behavior into specialized modules, related to scope detection, intent recognition, memory, inner speech, query generation, and outer speech, and (ii) exposes explicit policies for information sufficiency, execution readiness, and tool grounding. A dedicated memory agent maintains a bounded recent-history buffer, a compact core memory, and an archival store with semantic retrieval, coupled through controlled consolidation and revision policies. Models inspired by the notion of inner speech in cognitive theories provide a control-oriented internal textual flow that validates parameter completeness and triggers clarification before grounding, while a faithfulness constraint ties robot-to-human claims to an evidence bundle combining working context and retrieved tool outputs. We evaluate JANUS through module-level unit tests in a dietary assistance domain grounded on a knowledge graph, reporting high agreement with curated references and practical latency profiles. These results support factored reasoning as a promising path to scalable, auditable, and evidence-grounded robot assistance over extended interaction horizons.", "published": "2026-01-31", "updated": "2026-01-31", "authors": [ "Valerio Belcamino", "Mariya Kilina", "Alessandro Carfì", "Valeria Seidita", "Fulvio Mastrogiovanni", "Antonio Chella" ], "categories": [ "cs.RO", "cs.HC" ], "url": "https://arxiv.org/abs/2602.00675", "pdf_url": "https://arxiv.org/pdf/2602.00675", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2602.01869", "title": "Skill-Pro: Learning Reusable Skills from Experience via Non-Parametric PPO for LLM Agents", "abstract": "LLM-driven agents excel at sequential decision-making but often rely on on-the-fly reasoning, re-deriving solutions even in recurring scenarios. This insufficient experience reuse leads to computational redundancy and instability. To bridge this gap, we propose Skill-Pro, a framework enabling agents to autonomously learn reusable procedural skills from interaction experiences without parameter updates. By formalizing a Skill-MDP, Skill-Pro transforms passive episodic narratives into executable Skills defined by activation, execution, and termination conditions to ensure executability. To achieve reliable reusability without capability degradation, we introduce Non-Parametric PPO, which leverages semantic gradients for high-quality candidate generation and a PPO Gate for robust Skill verification. Through score-based maintenance, Skill-Pro sustains compact, high-quality procedural memory. Experimental results across in-domain, cross-task, and cross-agent scenarios demonstrate that Skill-Pro achieves superior reuse rates and significant gains with extreme memory compression. Visualized evolutionary trajectories and Skill distributions further reveal how Skill-Pro transparently accumulates, refines, and reuses procedural knowledge to facilitate long-term autonomy.", "published": "2026-02-02", "updated": "2026-05-28", "authors": [ "Qirui Mi", "Zhijian Ma", "Mengyue Yang", "Haoxuan Li", "Yisen Wang", "Haifeng Zhang", "Jun Wang" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2602.01869", "pdf_url": "https://arxiv.org/pdf/2602.01869", "matched_queries": [ "procedural-memory" ] }, { "arxiv_id": "2602.02007", "title": "Beyond RAG for Agent Memory: Retrieval by Decoupling and Aggregation", "abstract": "Standard Retrieval Augmented Generation (RAG) is poorly matched to agent memory. Unlike large heterogeneous corpora, agent memory forms a bounded and coherent interaction stream in which many spans are highly correlated or near duplicates. As a result, flat top-$k$ similarity retrieval often returns redundant context, while summary-centric hierarchies can blur the subtle details that distinguish one candidate from another. We argue that agent memory should follow the principle of decoupling before aggregation: the system should first isolate reusable facts, updates, and distinguishing details from similar histories, and only then organise them for efficient retrieval. Based on this principle, we propose xMemory, which constructs a revisable hierarchical memory structure from original messages to segments, memory components, and groups. xMemory segments interaction history into local events, decouples each segment into memory components, aggregates related components into high-level groups using a sparsity--semantic faithfulness objective, and maintains this structure incrementally as memory evolves. At inference time, xMemory retrieves top-down, first selecting a compact backbone of complementary groups and components, and then expanding to segments and raw messages only when additional evidence reduces the reader's uncertainty. Experiments on LoCoMo and PerLTQA across diverse open source and closed source LLMs show consistent gains in answer quality and inference token efficiency, supported by analyses of redundancy, evidence density, and coverage.", "published": "2026-02-02", "updated": "2026-05-12", "authors": [ "Zhanghao Hu", "Qinglin Zhu", "Runcong Zhao", "Di Liang", "Hanqi Yan", "Yulan He", "Lin Gui" ], "categories": [ "cs.CL", "cs.AI" ], "url": "https://arxiv.org/abs/2602.02007", "pdf_url": "https://arxiv.org/pdf/2602.02007", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2602.02164", "title": "Co-RedTeam: Orchestrated Security Discovery and Exploitation with LLM Agents", "abstract": "Large language models (LLMs) have shown promise in assisting cybersecurity tasks, yet existing approaches struggle with automatic vulnerability discovery and exploitation due to limited interaction, weak execution grounding, and a lack of experience reuse. We propose Co-RedTeam, a security-aware multi-agent framework designed to mirror real-world red-teaming workflows by integrating security-domain knowledge, code-aware analysis, execution-grounded iterative reasoning, and long-term memory. Co-RedTeam decomposes vulnerability analysis into coordinated discovery and exploitation stages, enabling agents to plan, execute, validate, and refine actions based on real execution feedback while learning from prior trajectories. Extensive evaluations on challenging security benchmarks demonstrate that Co-RedTeam consistently outperforms strong baselines across diverse backbone models, achieving over 60% success rate in vulnerability exploitation and over 10% absolute improvement in vulnerability detection. Ablation and iteration studies further confirm the critical role of execution feedback, structured interaction, and memory for building robust and generalizable cybersecurity agents.", "published": "2026-02-02", "updated": "2026-02-03", "authors": [ "Pengfei He", "Ash Fox", "Lesly Miculicich", "Stefan Friedli", "Daniel Fabian", "Burak Gokturk", "Jiliang Tang", "Chen-Yu Lee", "Tomas Pfister", "Long T. Le" ], "categories": [ "cs.LG", "cs.CR" ], "url": "https://arxiv.org/abs/2602.02164", "pdf_url": "https://arxiv.org/pdf/2602.02164", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2602.02369", "title": "Live-Evo: Online Evolution of Agentic Memory from Continuous Feedback", "abstract": "Large language model (LLM) agents are increasingly equipped with memory, which are stored experience and reusable guidance that can improve task-solving performance. Recent \\emph{self-evolving} systems update memory based on interaction outcomes, but most existing evolution pipelines are developed for static train/test splits and only approximate online learning by folding static benchmarks, making them brittle under true distribution shift and continuous feedback. We introduce \\textsc{Live-Evo}, an online self-evolving memory system that learns from a stream of incoming data over time. \\textsc{Live-Evo} decouples \\emph{what happened} from \\emph{how to use it} via an Experience Bank and a Meta-Guideline Bank, compiling task-adaptive guidelines from retrieved experiences for each task. To manage memory online, \\textsc{Live-Evo} maintains experience weights and updates them from feedback: experiences that consistently help are reinforced and retrieved more often, while misleading or stale experiences are down-weighted and gradually forgotten, analogous to reinforcement and decay in human memory. On the live \\textit{Prophet Arena} benchmark over a 10-week horizon, \\textsc{Live-Evo} improves Brier score by 20.8\\% and increases market returns by 12.9\\%, while also transferring to deep-research benchmarks with consistent gains over strong baselines. Our code is available at https://github.com/ag2ai/Live-Evo.", "published": "2026-02-02", "updated": "2026-02-02", "authors": [ "Yaolun Zhang", "Yiran Wu", "Yijiong Yu", "Qingyun Wu", "Huazheng Wang" ], "categories": [ "cs.AI", "cs.LG" ], "url": "https://arxiv.org/abs/2602.02369", "pdf_url": "https://arxiv.org/pdf/2602.02369", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2602.02474", "title": "MemSkill: Learning and Evolving Memory Skills for Self-Evolving Agents", "abstract": "Most Large Language Model (LLM) agent memory systems rely on a small set of static, hand-designed operations for extracting memory. These fixed procedures hard-code human priors about what to store and how to revise memory, making them rigid under diverse interaction patterns and inefficient on long histories. To this end, we present \\textbf{MemSkill}, which reframes these operations as learnable and evolvable memory skills, structured and reusable routines for extracting, consolidating, and pruning information from interaction traces. Inspired by the design philosophy of agent skills, MemSkill employs a \\emph{controller} that learns to select a small set of relevant skills, paired with an LLM-based \\emph{executor} that produces skill-guided memories. Beyond learning skill selection, MemSkill introduces a \\emph{designer} that periodically reviews hard cases where selected skills yield incorrect or incomplete memories, and evolves the skill set by proposing refinements and new skills. Together, MemSkill forms a closed-loop procedure that improves both the skill-selection policy and the skill set itself. Experiments on LoCoMo, LongMemEval, HotpotQA, and ALFWorld demonstrate that MemSkill improves task performance over strong baselines and generalizes well across settings. Further analyses shed light on how skills evolve, offering insights toward more adaptive, self-evolving memory management for LLM agents.", "published": "2026-02-02", "updated": "2026-05-24", "authors": [ "Haozhen Zhang", "Quanyu Long", "Jianzhu Bao", "Tao Feng", "Weizhi Zhang", "Haodong Yue", "Wenya Wang" ], "categories": [ "cs.CL", "cs.AI", "cs.LG" ], "url": "https://arxiv.org/abs/2602.02474", "pdf_url": "https://arxiv.org/pdf/2602.02474", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2602.03036", "title": "LatentMem: Customizing Latent Memory for Multi-Agent Systems", "abstract": "Large language model (LLM)-powered multi-agent systems (MAS) demonstrate remarkable collective intelligence, wherein multi-agent memory serves as a pivotal mechanism for continual adaptation. However, existing multi-agent memory designs remain constrained by two fundamental bottlenecks: (i) memory homogenization arising from the absence of role-aware customization, and (ii) information overload induced by excessively fine-grained memory entries. To address these limitations, we propose LatentMem, a learnable multi-agent memory framework designed to customize agent-specific memories in a token-efficient manner. Specifically, LatentMem comprises an experience bank that stores raw interaction trajectories in a lightweight form, and a memory composer that synthesizes compact latent memories conditioned on retrieved experience and agent-specific contexts. Further, we introduce Latent Memory Policy Optimization (LMPO), which propagates task-level optimization signals through latent memories to the composer, encouraging it to produce compact and high-utility representations. Extensive experiments across diverse benchmarks and mainstream MAS frameworks show that LatentMem achieves a performance gain of up to $19.36$% over vanilla settings and consistently outperforms existing memory architectures, without requiring any modifications to the underlying frameworks.", "published": "2026-02-03", "updated": "2026-03-09", "authors": [ "Muxin Fu", "Xiangyuan Xue", "Yafu Li", "Zefeng He", "Siyuan Huang", "Xiaoye Qu", "Yu Cheng", "Yang Yang" ], "categories": [ "cs.CL", "cs.LG", "cs.MA" ], "url": "https://arxiv.org/abs/2602.03036", "pdf_url": "https://arxiv.org/pdf/2602.03036", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2602.03224", "title": "TAME: A Trustworthy Test-Time Evolution of Agent Memory with Systematic Benchmarking", "abstract": "Test-time evolution of agent memory represents a pivotal paradigm for advancing AGI, as it strengthens complex reasoning through experience accumulation without requiring parameter updates. However, even during benign task evolution, agent safety alignment remains vulnerable, a phenomenon known as Agent Memory Misevolution. To evaluate this phenomenon, we construct the Trust-Memevo benchmark and find that agents exhibit an overall decline in trustworthiness across multiple tasks during benign task evolution. To address this issue, we propose TAME, a trust-aware memory evolution framework in which a shared memory bank is jointly governed by an Executor and an Evaluator. The Executor retrieves and applies transferable experiences to support task solving, while the Evaluator assesses the contribution of each utilized experience to the outcome and produces trust-aware feedback to guide subsequent memory use. This executor-evaluator loop enables memory to be selectively reinforced, cautiously reused, and continuously expanded over time. Experiments show that TAME mitigates memory misevolution while achieving strong task performance. In particular, on the GPT-5.2 AIME benchmark, TAME improves accuracy by 14.6 percentage points over the strongest existing method and maintains competitive trustworthiness.", "published": "2026-02-03", "updated": "2026-06-06", "authors": [ "Yu Cheng", "Yongkang Hu", "Jiuan Zhou", "Yushuo Zhang", "Yihang Chen", "Huichi Zhou", "Mingang Chen", "Zhizhong Zhang", "Kun Shao", "Yuan Xie", "Zhaoxia Yin" ], "categories": [ "cs.AI", "cs.LG" ], "url": "https://arxiv.org/abs/2602.03224", "pdf_url": "https://arxiv.org/pdf/2602.03224", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2602.03315", "title": "Memora: A Harmonic Memory Representation Balancing Abstraction and Specificity", "abstract": "Agent memory systems must accommodate continuously growing information while supporting efficient, context-aware retrieval for downstream tasks. Abstraction is essential for scaling agent memory, yet it often comes at the cost of specificity, obscuring the fine-grained details required for effective reasoning. We introduce Memora, a harmonic memory representation that structurally balances abstraction and specificity. Memora organizes information via its primary abstractions that index concrete memory values and consolidate related updates into unified memory entries, while cue anchors expand retrieval access across diverse aspects of the memory and connect related memories. Building on this structure, we employ a retrieval policy that actively exploits these memory connections to retrieve relevant information beyond direct semantic similarity. Theoretically, we show that standard Retrieval-Augmented Generation (RAG) and Knowledge Graph (KG)-based memory systems emerge as special cases of our framework. Empirically, Memora establishes a new state-of-the-art on the LoCoMo and LongMemEval benchmarks, demonstrating better retrieval relevance and reasoning effectiveness as memory scales.", "published": "2026-02-03", "updated": "2026-07-02", "authors": [ "Menglin Xia", "Xuchao Zhang", "Shantanu Dixit", "Paramaguru Harimurugan", "Rujia Wang", "Victor Ruhle", "Robert Sim", "Chetan Bansal", "Saravan Rajmohan" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2602.03315", "pdf_url": "https://arxiv.org/pdf/2602.03315", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2602.04482", "title": "ProAgentBench: Evaluating LLM Agents for Proactive Assistance with Real-World Data", "abstract": "Proactive agents that anticipate user intentions without explicit prompts represent a significant evolution in human-AI interaction, promising to reduce cognitive load and streamline workflows. However, existing datasets suffer from two critical deficiencies: (1) reliance on LLM-synthesized data that fails to capture authentic human decision-making patterns, and (2) focus on isolated tasks rather than continuous workflows, missing the pre-assistance behavioral context essential for learning proactive intervention signals. To address these gaps, we introduce ProAgentBench, a rigorous benchmark for proactive agents in working scenarios. Our contributions include: (1) a hierarchical task framework that decomposes proactive assistance into timing prediction and assist content generation; (2) a privacy-compliant dataset with 28,000+ events from 500+ hours of real user sessions, preserving bursty interaction patterns (burstiness B=0.787) absent in synthetic data; and (3) extensive experiments that evaluates LLM- and VLM-based baselines. Numerically, we showed that long-term memory and historical context significantly enhance prediction accuracy, while real-world training data substantially outperforms synthetic alternatives. We release our dataset and code at https://anonymous.4open.science/r/ProAgentBench-6BC0.", "published": "2026-02-04", "updated": "2026-02-09", "authors": [ "Yuanbo Tang", "Huaze Tang", "Tingyu Cao", "Lam Nguyen", "Anping Zhang", "Xinwen Cao", "Chunkang Liu", "Wenbo Ding", "Yang Li" ], "categories": [ "cs.HC" ], "url": "https://arxiv.org/abs/2602.04482", "pdf_url": "https://arxiv.org/pdf/2602.04482", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2602.04640", "title": "Towards Structured, State-Aware, and Execution-Grounded Reasoning for Software Engineering Agents", "abstract": "Software Engineering (SE) agents have shown promising abilities in supporting various SE tasks. Current SE agents remain fundamentally reactive, making decisions mainly based on conversation history and the most recent response. However, this reactive design provides no explicit structure or persistent state within the agent's memory, making long-horizon reasoning challenging. As a result, SE agents struggle to maintain a coherent understanding across reasoning steps, adapt their hypotheses as new evidence emerges, or incorporate execution feedback into the mental reasoning model of the system state. In this position paper, we argue that, to further advance SE agents, we need to move beyond reactive behavior toward a structured, state-aware, and execution-grounded reasoning. We outline how explicit structure, persistent and evolving state, and the integration of execution-grounded feedback can help SE agents perform more coherent and reliable reasoning in long-horizon tasks. We also provide an initial roadmap for developing next-generation SE agents that can more effectively perform real-world tasks.", "published": "2026-02-04", "updated": "2026-02-04", "authors": [ "Tse-Hsun", "Chen" ], "categories": [ "cs.SE", "cs.AI" ], "url": "https://arxiv.org/abs/2602.04640", "pdf_url": "https://arxiv.org/pdf/2602.04640", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2603.00026", "title": "ActMem: Bridging the Gap Between Memory Retrieval and Reasoning in LLM Agents", "abstract": "Memory management is essential for LLM agents in long-term interactions. Current memory frameworks typically treat agents as passive ``recorders'' and retrieve information without understanding its deeper implications. They may fail in scenarios requiring reasoning and complex decision-making. To bridge this critical gap, we propose a novel actionable memory framework called ActMem that integrates memory retrieval with active causal reasoning. ActMem transforms unstructured dialogue history into a structured causal and semantic graph. By leveraging counterfactual reasoning and commonsense completion, it enables agents to deduce implicit constraints and resolve potential conflicts between past states and current intentions. Furthermore, we introduce a comprehensive dataset ActMemEval to evaluate agent reasoning capabilities in logic-driven scenarios, moving beyond the fact-retrieval focus of existing memory benchmarks. Experiments demonstrate that ActMem significantly outperforms baselines in handling complex, memory-dependent tasks, paving the way for more consistent and reliable intelligent assistants.", "published": "2026-02-04", "updated": "2026-06-17", "authors": [ "Xiaohui Zhang", "Zequn Sun", "Chengyuan Yang", "Yaqin Jin", "Yazhong Zhang", "Wei Hu" ], "categories": [ "cs.CL", "cs.AI", "cs.IR" ], "url": "https://arxiv.org/abs/2603.00026", "pdf_url": "https://arxiv.org/pdf/2603.00026", "matched_queries": [ "memory-evaluation" ] }, { "arxiv_id": "2602.05665", "title": "Graph-based Agent Memory: Taxonomy, Techniques, and Applications", "abstract": "Memory emerges as the core module in the Large Language Model (LLM)-based agents for long-horizon complex tasks (e.g., multi-turn dialogue, game playing, scientific discovery), where memory can enable knowledge accumulation, iterative reasoning and self-evolution. Among diverse paradigms, graph stands out as a powerful structure for agent memory due to the intrinsic capabilities to model relational dependencies, organize hierarchical information, and support efficient retrieval. This survey presents a comprehensive review of agent memory from the graph-based perspective. First, we introduce a taxonomy of agent memory, including short-term vs. long-term memory, knowledge vs. experience memory, non-structural vs. structural memory, with an implementation view of graph-based memory. Second, according to the life cycle of agent memory, we systematically analyze the key techniques in graph-based agent memory, covering memory extraction for transforming the data into the contents, storage for organizing the data efficiently, retrieval for retrieving the relevant contents from memory to support reasoning, and evolution for updating the contents in the memory. Third, we summarize the open-sourced libraries and benchmarks that support the development and evaluation of self-evolving agent memory. We also explore diverse application scenarios. Finally, we identify critical challenges and future research directions. This survey aims to offer actionable insights to advance the development of more efficient and reliable graph-based agent memory systems. All the related resources, including research papers, open-source data, and projects, are collected for the community in https://github.com/DEEP-PolyU/Awesome-GraphMemory.", "published": "2026-02-05", "updated": "2026-02-05", "authors": [ "Chang Yang", "Chuang Zhou", "Yilin Xiao", "Su Dong", "Luyao Zhuang", "Yujing Zhang", "Zhu Wang", "Zijin Hong", "Zheng Yuan", "Zhishang Xiang", "Shengyuan Chen", "Huachi Zhou", "Qinggang Zhang", "Ninghao Liu", "Jinsong Su", "Xinrun Wang", "Yi Chang", "Xiao Huang" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2602.05665", "pdf_url": "https://arxiv.org/pdf/2602.05665", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2602.06025", "title": "Learning Query-Aware Budget-Tier Routing for Runtime Agent Memory", "abstract": "Memory is increasingly central to Large Language Model (LLM) agents operating beyond a single context window, yet most existing systems rely on offline, query-agnostic memory construction that can be inefficient and may discard query-critical information. Although runtime memory utilization is a natural alternative, prior work often incurs substantial overhead and offers limited explicit control over the performance-cost trade-off. In this work, we present \\textbf{BudgetMem}, a runtime agent memory framework for explicit, query-aware performance-cost control. BudgetMem structures memory processing as a set of memory modules, each offered in three budget tiers (i.e., \\textsc{Low}/\\textsc{Mid}/\\textsc{High}). A lightweight router performs budget-tier routing across modules to balance task performance and memory construction cost, which is implemented as a compact neural policy trained with reinforcement learning. Using BudgetMem as a unified testbed, we study three complementary strategies for realizing budget tiers: implementation (method complexity), reasoning (inference behavior), and capacity (module model size). Across LoCoMo, LongMemEval, and HotpotQA, BudgetMem surpasses strong baselines when performance is prioritized (i.e., high-budget setting), and delivers better accuracy-cost frontiers under tighter budgets. Moreover, our analysis disentangles the strengths and weaknesses of different tiering strategies, clarifying when each axis delivers the most favorable trade-offs under varying budget regimes.", "published": "2026-02-05", "updated": "2026-05-27", "authors": [ "Haozhen Zhang", "Haodong Yue", "Tao Feng", "Quanyu Long", "Jianzhu Bao", "Bowen Jin", "Weizhi Zhang", "Xiao Li", "Jiaxuan You", "Chengwei Qin", "Wenya Wang" ], "categories": [ "cs.CL", "cs.AI", "cs.LG" ], "url": "https://arxiv.org/abs/2602.06025", "pdf_url": "https://arxiv.org/pdf/2602.06025", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2602.17692", "title": "Agentic Unlearning: When LLM Agent Meets Machine Unlearning", "abstract": "In this paper, we introduce \\textbf{agentic unlearning} which removes specified information from both model parameters and persistent memory in agents with closed-loop interaction. Existing unlearning methods target parameters alone, leaving two critical gaps: (i) parameter-memory backflow, where retrieval reactivates parametric remnants or memory artifacts reintroduce sensitive content, and (ii) the absence of a unified strategy that covers both parameter and memory pathways. We present Synchronized Backflow Unlearning (SBU), a framework that unlearns jointly across parameter and memory pathways. The memory pathway performs dependency closure-based unlearning that prunes isolated entities while logically invalidating shared artifacts. The parameter pathway employs stochastic reference alignment to guide model outputs toward a high-entropy prior. These pathways are integrated via a synchronized dual-update protocol, forming a closed-loop mechanism where memory unlearning and parametric suppression reinforce each other to prevent cross-pathway recontamination. Experiments on medical QA benchmarks show that SBU reduces traces of targeted private information across both pathways with limited degradation on retained data.", "published": "2026-02-06", "updated": "2026-02-28", "authors": [ "Bin Wang", "Fan Wang", "Pingping Wang", "Jinyu Cong", "Yang Yu", "Yilong Yin", "Zhongyi Han", "Benzheng Wei" ], "categories": [ "cs.LG", "cs.AI" ], "url": "https://arxiv.org/abs/2602.17692", "pdf_url": "https://arxiv.org/pdf/2602.17692", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2603.03296", "title": "PlugMem: A Task-Agnostic Plugin Memory Module for LLM Agents", "abstract": "Long-term memory is essential for large language model (LLM) agents operating in complex environments, yet existing memory designs are either task-specific and non-transferable, or task-agnostic but less effective due to low task-relevance and context explosion from raw memory retrieval. We propose PlugMem, a task-agnostic plugin memory module that can be attached to arbitrary LLM agents without task-specific redesign. Motivated by the fact that decision-relevant information is concentrated as abstract knowledge rather than raw experience, we draw on cognitive science to structure episodic memories into a compact, extensible knowledge-centric memory graph that explicitly represents propositional and prescriptive knowledge. This representation enables efficient memory retrieval and reasoning over task-relevant knowledge, rather than verbose raw trajectories, and departs from other graph-based methods like GraphRAG by treating knowledge as the unit of memory access and organization instead of entities or text chunks. We evaluate PlugMem unchanged across three heterogeneous benchmarks (long-horizon conversational question answering, multi-hop knowledge retrieval, and web agent tasks). The results show that PlugMem consistently outperforms task-agnostic baselines and exceeds task-specific memory designs, while also achieving the highest information density under a unified information-theoretic analysis. Code and data are available at https://github.com/TIMAN-group/PlugMem.", "published": "2026-02-06", "updated": "2026-02-06", "authors": [ "Ke Yang", "Zixi Chen", "Xuan He", "Jize Jiang", "Michel Galley", "Chenglong Wang", "Jianfeng Gao", "Jiawei Han", "ChengXiang Zhai" ], "categories": [ "cs.CL", "cs.AI", "cs.IR" ], "url": "https://arxiv.org/abs/2603.03296", "pdf_url": "https://arxiv.org/pdf/2603.03296", "matched_queries": [ "episodic-memory", "long-term-memory" ] }, { "arxiv_id": "2602.07398", "title": "AgentSys: Secure and Dynamic LLM Agents Through Explicit Hierarchical Memory Management", "abstract": "Indirect prompt injection threatens LLM agents by embedding malicious instructions in external content, enabling unauthorized actions and data theft. LLM agents maintain working memory through their context window, which stores interaction history for decision-making. Conventional agents indiscriminately accumulate all tool outputs and reasoning traces in this memory, creating two critical vulnerabilities: (1) injected instructions persist throughout the workflow, granting attackers multiple opportunities to manipulate behavior, and (2) verbose, non-essential content degrades decision-making capabilities. Existing defenses treat bloated memory as given and focus on remaining resilient, rather than reducing unnecessary accumulation to prevent the attack. We present AgentSys, a framework that defends against indirect prompt injection through explicit memory management. Inspired by process memory isolation in operating systems, AgentSys organizes agents hierarchically: a main agent spawns worker agents for tool calls, each running in an isolated context and able to spawn nested workers for subtasks. External data and subtask traces never enter the main agent's memory; only schema-validated return values can cross boundaries through deterministic JSON parsing. Ablations show isolation alone cuts attack success to 2.19%, and adding a validator/sanitizer further improves defense with event-triggered checks whose overhead scales with operations rather than context length. On AgentDojo and ASB, AgentSys achieves 0.78% and 4.25% attack success while slightly improving benign utility over undefended baselines. It remains robust to adaptive attackers and across multiple foundation models, showing that explicit memory management enables secure, dynamic LLM agent architectures. Our code is available at: https://github.com/ruoyaow/agentsys-memory.", "published": "2026-02-07", "updated": "2026-02-07", "authors": [ "Ruoyao Wen", "Hao Li", "Chaowei Xiao", "Ning Zhang" ], "categories": [ "cs.CR", "cs.AI" ], "url": "https://arxiv.org/abs/2602.07398", "pdf_url": "https://arxiv.org/pdf/2602.07398", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2602.07624", "title": "M2A: Multimodal Memory Agent with Dual-Layer Hybrid Memory for Long-Term Personalized Interactions", "abstract": "This work addresses the challenge of personalized question answering in long-term human-machine interactions: when conversational history spans weeks or months and exceeds the context window, existing personalization mechanisms struggle to continuously absorb and leverage users' incremental concepts, aliases, and preferences. Current personalized multimodal models are predominantly static-concepts are fixed at initialization and cannot evolve during interactions. We propose M2A, an agentic dual-layer hybrid memory system that maintains personalized multimodal information through online updates. The system employs two collaborative agents: ChatAgent manages user interactions and autonomously decides when to query or update memory, while MemoryManager breaks down memory requests from ChatAgent into detailed operations on the dual-layer memory bank, which couples a RawMessageStore (immutable conversation log) with a SemanticMemoryStore (high-level observations), providing memories at different granularities. In addition, we develop a reusable data synthesis pipeline that injects concept-grounded sessions from Yo'LLaVA and MC-LLaVA into LoCoMo long conversations while preserving temporal coherence. Experiments show that M2A significantly outperforms baselines, demonstrating that transforming personalization from one-shot configuration to a co-evolving memory mechanism provides a viable path for high-quality individualized responses in long-term multimodal interactions. The code is available at https://github.com/Little-Fridge/M2A.", "published": "2026-02-07", "updated": "2026-02-07", "authors": [ "Junyu Feng", "Binxiao Xu", "Jiayi Chen", "Mengyu Dai", "Cenyang Wu", "Haodong Li", "Bohan Zeng", "Yunliu Xie", "Hao Liang", "Ming Lu", "Wentao Zhang" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2602.07624", "pdf_url": "https://arxiv.org/pdf/2602.07624", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2602.07755", "title": "Learning to Continually Learn via Meta-learning Agentic Memory Designs", "abstract": "The statelessness of foundation models bottlenecks agentic systems' ability to continually learn, a core capability for long-horizon reasoning and adaptation. To address this limitation, agentic systems commonly incorporate memory modules to retain and reuse past experience, aiming for continual learning during test time. However, most existing memory designs are human-crafted and fixed, which limits their ability to adapt to the diversity and non-stationarity of real-world tasks. In this paper, we introduce ALMA (Automated meta-Learning of Memory designs for Agentic systems), a framework that meta-learns memory designs to replace hand-engineered memory designs, therefore minimizing human effort and enabling agentic systems to be continual learners across diverse domains. Our approach employs a Meta Agent that searches over memory designs expressed as executable code in an open-ended manner, theoretically allowing the discovery of arbitrary memory designs, including database schemas as well as their retrieval and update mechanisms. Extensive experiments across four sequential decision-making domains demonstrate that the learned memory designs enable more effective and efficient learning from experience than state-of-the-art human-crafted memory designs on all benchmarks. When developed and deployed safely, ALMA represents a step toward self-improving AI systems that learn to be adaptive, continual learners.", "published": "2026-02-08", "updated": "2026-02-08", "authors": [ "Yiming Xiong", "Shengran Hu", "Jeff Clune" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2602.07755", "pdf_url": "https://arxiv.org/pdf/2602.07755", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2602.07885", "title": "MemFly: On-the-Fly Memory Optimization via Information Bottleneck", "abstract": "Long-term memory enables large language model agents to tackle complex tasks through historical interactions. However, existing frameworks encounter a fundamental dilemma between compressing redundant information efficiently and maintaining precise retrieval for downstream tasks. To bridge this gap, we propose MemFly, a framework grounded in information bottleneck principles that facilitates on-the-fly memory evolution for LLMs. Our approach minimizes compression entropy while maximizing relevance entropy via a gradient-free optimizer, constructing a stratified memory structure for efficient storage. To fully leverage MemFly, we develop a hybrid retrieval mechanism that seamlessly integrates semantic, symbolic, and topological pathways, incorporating iterative refinement to handle complex multi-hop queries. Comprehensive experiments demonstrate that MemFly substantially outperforms state-of-the-art baselines in memory coherence, response fidelity, and accuracy.", "published": "2026-02-08", "updated": "2026-02-08", "authors": [ "Zhenyuan Zhang", "Xianzhang Jia", "Zhiqin Yang", "Zhenbo Song", "Wei Xue", "Sirui Han", "Yike Guo" ], "categories": [ "cs.AI", "cs.LG" ], "url": "https://arxiv.org/abs/2602.07885", "pdf_url": "https://arxiv.org/pdf/2602.07885", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2602.08369", "title": "MemAdapter: Fast Alignment across Agent Memory Paradigms via Generative Subgraph Retrieval", "abstract": "Memory mechanism is a core component of LLM-based agents, enabling reasoning and knowledge discovery over long-horizon contexts. Existing agent memory systems are typically designed within isolated paradigms (e.g., explicit, parametric, or latent memory) with tightly coupled retrieval methods that hinder cross-paradigm generalization and fusion. In this work, we take a first step toward unifying heterogeneous memory paradigms within a single memory system. We propose MemAdapter, a memory retrieval framework that enables fast alignment across agent memory paradigms. MemAdapter adopts a two-stage training strategy: (1) training a generative subgraph retriever from the unified memory space, and (2) adapting the retriever to unseen memory paradigms by training a lightweight alignment module through contrastive learning. This design improves the flexibility for memory retrieval and substantially reduces alignment cost across paradigms. Comprehensive experiments on three public evaluation benchmarks demonstrate that the generative subgraph retriever consistently outperforms five strong agent memory systems across three memory paradigms and agent model scales. Notably, MemAdapter completes cross-paradigm alignment within 13 minutes on a single GPU, achieving superior performance over original memory retrievers with less than 5% of training compute. Furthermore, MemAdapter enables effective zero-shot fusion across memory paradigms, highlighting its potential as a plug-and-play solution for agent memory systems.", "published": "2026-02-09", "updated": "2026-02-09", "authors": [ "Xin Zhang", "Kailai Yang", "Chenyue Li", "Hao Li", "Qiyu Wei", "Jun'ichi Tsujii", "Sophia Ananiadou" ], "categories": [ "cs.AI", "cs.CL", "cs.LG" ], "url": "https://arxiv.org/abs/2602.08369", "pdf_url": "https://arxiv.org/pdf/2602.08369", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2602.09319", "title": "Benchmarking Knowledge-Extraction Attack and Defense on Retrieval-Augmented Generation", "abstract": "Retrieval-Augmented Generation (RAG) has become a cornerstone of knowledge-intensive applications, including enterprise chatbots, healthcare assistants, and agentic memory management. However, recent studies show that knowledge-extraction attacks can recover sensitive knowledge-base content through maliciously crafted queries, raising serious intellectual property and privacy concerns. While prior work has explored individual attack and defense techniques, the research landscape remains fragmented, spanning heterogeneous retrieval embeddings, diverse generation models, and evaluations based on non-standardized metrics and inconsistent datasets. To address this gap, we introduce the first systematic benchmark for knowledge-extraction attacks on RAG systems. Our benchmark covers broad attack/defense strategies, representative retrieval embedding models, open/closed-source generators, (non) graph-based indexing, all evaluated under a unified experimental framework with standardized protocols across multiple datasets spanning diverse languages. By consolidating the experimental landscape and enabling reproducible, comparable evaluation, this benchmark provides actionable insights and a practical foundation for developing privacy-preserving RAG systems in the face of emerging knowledge extraction threats.", "published": "2026-02-10", "updated": "2026-06-08", "authors": [ "Zhisheng Qi", "Utkarsh Sahu", "Li Ma", "Haoyu Han", "Ryan Rossi", "Franck Dernoncourt", "Mahantesh Halappanavar", "Nesreen Ahmed", "Yushun Dong", "Yue Zhao", "Yu Zhang", "Yu Wang" ], "categories": [ "cs.CR" ], "url": "https://arxiv.org/abs/2602.09319", "pdf_url": "https://arxiv.org/pdf/2602.09319", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2602.09712", "title": "TraceMem: Weaving Narrative Memory Schemata from User Conversational Traces", "abstract": "Sustaining long-term interactions remains a bottleneck for Large Language Models (LLMs), as their limited context windows struggle to manage dialogue histories that extend over time. Existing memory systems often treat interactions as disjointed snippets, failing to capture the underlying narrative coherence of the dialogue stream. We propose TraceMem, a cognitively-inspired framework that weaves structured, narrative memory schemata from user conversational traces through a three-stage pipeline: (1) Short-term Memory Processing, which employs a deductive topic segmentation approach to demarcate episode boundaries and extract semantic representation; (2) Synaptic Memory Consolidation, a process that summarizes episodes into episodic memories before distilling them alongside semantics into user-specific traces; and (3) Systems Memory Consolidation, which utilizes two-stage hierarchical clustering to organize these traces into coherent, time-evolving narrative threads under unifying themes. These threads are encapsulated into structured user memory cards, forming narrative memory schemata. For memory utilization, we provide an agentic search mechanism to enhance reasoning process. Evaluation on the LoCoMo benchmark shows that TraceMem achieves state-of-the-art performance with a brain-inspired architecture. Analysis shows that by constructing coherent narratives, it surpasses baselines in multi-hop and temporal reasoning, underscoring its essential role in deep narrative comprehension. Additionally, we provide an open discussion on memory systems, offering our perspectives and future outlook on the field. Our code implementation is available at: https://github.com/YimingShu-teay/TraceMem", "published": "2026-02-10", "updated": "2026-02-10", "authors": [ "Yiming Shu", "Pei Liu", "Tiange Zhang", "Ruiyang Gao", "Jun Ma", "Chen Sun" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2602.09712", "pdf_url": "https://arxiv.org/pdf/2602.09712", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2602.10715", "title": "Locomo-Plus: Beyond-Factual Cognitive Memory Evaluation Framework for LLM Agents", "abstract": "Long-term conversational memory is a core capability for LLM-based dialogue systems, yet existing benchmarks and evaluation protocols primarily focus on surface-level factual recall. In realistic interactions, appropriate responses often depend on implicit constraints such as user state, goals, or values that are not explicitly queried later. To evaluate this setting, we introduce \\textbf{LoCoMo-Plus}, a benchmark for assessing cognitive memory under cue--trigger semantic disconnect, where models must retain and apply latent constraints across long conversational contexts. We further show that conventional string-matching metrics and explicit task-type prompting are misaligned with such scenarios, and propose a unified evaluation framework based on constraint consistency. Experiments across diverse backbone models, retrieval-based methods, and memory systems demonstrate that cognitive memory remains challenging and reveals failures not captured by existing benchmarks. Our code and evaluation framework are publicly available at: https://github.com/xjtuleeyf/Locomo-Plus.", "published": "2026-02-11", "updated": "2026-02-11", "authors": [ "Yifei Li", "Weidong Guo", "Lingling Zhang", "Rongman Xu", "Muye Huang", "Hui Liu", "Lijiao Xu", "Yu Xu", "Jun Liu" ], "categories": [ "cs.CL", "cs.AI" ], "url": "https://arxiv.org/abs/2602.10715", "pdf_url": "https://arxiv.org/pdf/2602.10715", "matched_queries": [ "memory-evaluation" ] }, { "arxiv_id": "2602.11243", "title": "Evaluating Memory Structure in LLM Agents", "abstract": "Modern LLM-based agents and chat assistants rely on long-term memory frameworks to store reusable knowledge, recall user preferences, and augment reasoning. As researchers create more complex memory architectures, it becomes increasingly difficult to analyze their capabilities and guide future memory designs. Most long-term memory benchmarks focus on simple fact retention, multi-hop recall, and time-based changes. While undoubtedly important, these capabilities can often be achieved with simple retrieval-augmented LLMs and do not test complex memory hierarchies. To bridge this gap, we propose StructMemEval - a benchmark that tests the agent's ability to organize its long-term memory, not just factual recall. We gather a suite of tasks that humans solve by organizing their knowledge in a specific structure: transaction ledgers, to-do lists, trees and others. Our initial experiments show that simple retrieval-augmented LLMs struggle with these tasks, whereas memory agents can reliably solve them if prompted how to organize their memory. However, we also find that modern LLMs do not always recognize the memory structure when not prompted to do so. This highlights an important direction for future improvements in both LLM training and memory frameworks.", "published": "2026-02-11", "updated": "2026-05-22", "authors": [ "Alina Shutova", "Alexandra Olenina", "Ivan Vinogradov", "Anton Sinitsin" ], "categories": [ "cs.LG", "cs.CL" ], "url": "https://arxiv.org/abs/2602.11243", "pdf_url": "https://arxiv.org/pdf/2602.11243", "matched_queries": [ "agent-memory", "long-term-memory", "memory-evaluation" ] }, { "arxiv_id": "2602.13530", "title": "REMem: Reasoning with Episodic Memory in Language Agent", "abstract": "Humans excel at remembering concrete experiences along spatiotemporal contexts and performing reasoning across those events, i.e., the capacity for episodic memory. In contrast, memory in language agents remains mainly semantic, and current agents are not yet capable of effectively recollecting and reasoning over interaction histories. We identify and formalize the core challenges of episodic recollection and reasoning from this gap, and observe that existing work often overlooks episodicity, lacks explicit event modeling, or overemphasizes simple retrieval rather than complex reasoning. We present REMem, a two-phase framework for constructing and reasoning with episodic memory: 1) Offline indexing, where REMem converts experiences into a hybrid memory graph that flexibly links time-aware gists and facts. 2) Online inference, where REMem employs an agentic retriever with carefully curated tools for iterative retrieval over the memory graph. Comprehensive evaluation across four episodic memory benchmarks shows that REMem substantially outperforms state-of-the-art memory systems such as Mem0 and HippoRAG 2, showing 3.4% and 13.4% absolute improvements on episodic recollection and reasoning tasks, respectively. Moreover, REMem also demonstrates more robust refusal behavior for unanswerable questions.", "published": "2026-02-13", "updated": "2026-02-28", "authors": [ "Yiheng Shu", "Saisri Padmaja Jonnalagedda", "Xiang Gao", "Bernal Jiménez Gutiérrez", "Weijian Qi", "Kamalika Das", "Huan Sun", "Yu Su" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2602.13530", "pdf_url": "https://arxiv.org/pdf/2602.13530", "matched_queries": [ "memory-evaluation" ] }, { "arxiv_id": "2602.18493", "title": "Learning to Remember: End-to-End Training of Memory Agents for Long-Context Reasoning", "abstract": "Long-context LLMs and Retrieval-Augmented Generation (RAG) systems process information passively, deferring state tracking, contradiction resolution, and evidence aggregation to query time, which becomes brittle under ultra long streams with frequent updates. We propose the Unified Memory Agent (UMA), an end-to-end reinforcement learning framework that unifies memory operations and question answering within a single policy. UMA maintains a dual memory representation: a compact core summary for global context and a structured Memory Bank that supports explicit CRUD (create, update, delete, reorganize) over key value entries, enabling proactive consolidation during streaming. To evaluate long-horizon memory behavior, we introduce Ledger-QA, a diagnostic benchmark for continuous state tracking where answers are latent values derived from accumulated updates rather than lo cal span retrieval. Across 13 datasets spanning Ledger-QA, Test-Time Learning, and Accurate Retrieval, UMA substantially outperforms long-context and RAG baselines on dynamic reasoning and learning tasks while remaining competitive on standard retrieval benchmarks, underscoring the importance of learned, end-to-end memory management.", "published": "2026-02-13", "updated": "2026-02-13", "authors": [ "Kehao Zhang", "Shangtong Gui", "Sheng Yang", "Wei Chen", "Yang Feng" ], "categories": [ "cs.LG", "cs.AI" ], "url": "https://arxiv.org/abs/2602.18493", "pdf_url": "https://arxiv.org/pdf/2602.18493", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2602.13594", "title": "Hippocampus: An Efficient and Scalable Memory Module for Agentic AI", "abstract": "Agentic AI require persistent memory to store user-specific histories beyond the limited context window of LLMs. Existing memory systems use dense vector databases or knowledge-graph traversal (or hybrid), incurring high retrieval latency and poor storage scalability. We introduce Hippocampus, an agentic memory management system that uses compact binary signatures for semantic search and lossless token-ID streams for exact content reconstruction. Its core is a Dynamic Wavelet Matrix (DWM) that compresses and co-indexes both streams to support ultra-fast search in the compressed domain, thus avoiding costly dense-vector or graph computations. This design scales linearly with memory size, making it suitable for long-horizon agentic deployments. Empirically, our evaluation shows that Hippocampus reduces end-to-end retrieval latency by up to 31$\\times$ and cuts per-query token footprint by up to 14$\\times$, while maintaining accuracy on both LoCoMo and LongMemEval benchmarks.", "published": "2026-02-14", "updated": "2026-02-14", "authors": [ "Yi Li", "Lianjie Cao", "Faraz Ahmed", "Puneet Sharma", "Bingzhe Li" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2602.13594", "pdf_url": "https://arxiv.org/pdf/2602.13594", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2602.13933", "title": "HyMem: Hybrid Memory Architecture with Dynamic Retrieval Scheduling", "abstract": "Large language model (LLM) agents demonstrate strong performance in short-text contexts but often underperform in extended dialogues due to inefficient memory management. Existing approaches face a fundamental trade-off between efficiency and effectiveness: memory compression risks losing critical details required for complex reasoning, while retaining raw text introduces unnecessary computational overhead for simple queries. The crux lies in the limitations of monolithic memory representations and static retrieval mechanisms, which fail to emulate the flexible and proactive memory scheduling capabilities observed in humans, thus struggling to adapt to diverse problem scenarios. Inspired by the principle of cognitive economy, we propose HyMem, a hybrid memory architecture that enables dynamic on-demand scheduling through multi-granular memory representations. HyMem adopts a dual-granular storage scheme paired with a dynamic two-tier retrieval system: a lightweight module constructs summary-level context for efficient response generation, while an LLM-based deep module is selectively activated only for complex queries, augmented by a reflection mechanism for iterative reasoning refinement. Experiments show that HyMem achieves strong performance on both the LOCOMO and LongMemEval benchmarks, outperforming full-context while reducing computational cost by 92.6\\%, establishing a state-of-the-art balance between efficiency and performance in long-term memory management.", "published": "2026-02-15", "updated": "2026-05-01", "authors": [ "Xiaochen Zhao", "Kaikai Wang", "Xiaowen Zhang", "Chen Yao", "Aili Wang" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2602.13933", "pdf_url": "https://arxiv.org/pdf/2602.13933", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2602.14038", "title": "Choosing How to Remember: Adaptive Memory Structures for LLM Agents", "abstract": "Memory is critical for enabling large language model (LLM) based agents to maintain coherent behavior over long-horizon interactions. However, existing agent memory systems suffer from two key gaps: they rely on a one-size-fits-all memory structure and do not model memory structure selection as a context-adaptive decision, limiting their ability to handle heterogeneous interaction patterns and resulting in suboptimal performance. We propose a unified framework, FluxMem, that enables adaptive memory organization for LLM agents. Our framework equips agents with multiple complementary memory structures. It explicitly learns to select among these structures based on interaction-level features, using offline supervision derived from downstream response quality and memory utilization. To support robust long-horizon memory evolution, we further introduce a three-level memory hierarchy and a Beta Mixture Model-based probabilistic gate for distribution-aware memory fusion, replacing brittle similarity thresholds. Experiments on two long-horizon benchmarks, PERSONAMEM and LoCoMo, demonstrate that our method achieves average improvements of 9.18% and 6.14%.", "published": "2026-02-15", "updated": "2026-02-15", "authors": [ "Mingfei Lu", "Mengjia Wu", "Feng Liu", "Jiawei Xu", "Weikai Li", "Haoyang Wang", "Zhengdong Hu", "Ying Ding", "Yizhou Sun", "Jie Lu", "Yi Zhang" ], "categories": [ "cs.AI", "cs.LG" ], "url": "https://arxiv.org/abs/2602.14038", "pdf_url": "https://arxiv.org/pdf/2602.14038", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2602.15274", "title": "When Remembering and Planning are Worth it: Navigating under Change", "abstract": "We explore how different types and uses of memory can aid spatial navigation in changing uncertain environments. In the simple foraging task we study, every day, our agent has to find its way from its home, through barriers, to food. Moreover, the world is non-stationary: from day to day, the location of the barriers and food may change, and the agent's sensing such as its location information is uncertain and very limited. Any model construction, such as a map, and use, such as planning, needs to be robust against these challenges, and if any learning is to be useful, it needs to be adequately fast. We look at a range of strategies, from simple to sophisticated, with various uses of memory and learning. We find that an architecture that can incorporate multiple strategies is required to handle (sub)tasks of a different nature, in particular for exploration and search, when food location is not known, and for planning a good path to a remembered (likely) food location. An agent that utilizes non-stationary probability learning techniques to keep updating its (episodic) memories and that uses those memories to build maps and plan on the fly (imperfect maps, i.e. noisy and limited to the agent's experience) can be increasingly and substantially more efficient than the simpler (minimal-memory) agents, as the task difficulties such as distance to goal are raised, as long as the uncertainty, from localization and change, is not too large.", "published": "2026-02-17", "updated": "2026-02-17", "authors": [ "Omid Madani", "J. Brian Burns", "Reza Eghbali", "Thomas L. Dean" ], "categories": [ "cs.AI", "cs.LG" ], "url": "https://arxiv.org/abs/2602.15274", "pdf_url": "https://arxiv.org/pdf/2602.15274", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2602.15513", "title": "HIMM: Human-Inspired Long-Term Memory Modeling for Embodied Exploration and Question Answering", "abstract": "Deploying Multimodal Large Language Models as the brain of embodied agents remains challenging, particularly under long-horizon observations and limited context budgets. Existing memory assisted methods often rely on textual summaries, which discard rich visual and spatial details and remain brittle in non-stationary environments. In this work, we propose a non-parametric memory framework that explicitly disentangles episodic and semantic memory for embodied exploration and question answering. Our retrieval-first, reasoning-assisted paradigm recalls episodic experiences via semantic similarity and verifies them through visual reasoning, enabling robust reuse of past observations without rigid geometric alignment. In parallel, we introduce a program-style rule extraction mechanism that converts experiences into structured, reusable semantic memory, facilitating cross-environment generalization. Extensive experiments demonstrate state-of-the-art performance on embodied question answering and exploration benchmarks, yielding a 7.3% gain in LLM-Match and an 11.4% gain in LLM MatchXSPL on A-EQA, as well as +7.7% success rate and +6.8% SPL on GOAT-Bench. Analyses reveal that our episodic memory primarily improves exploration efficiency, while semantic memory strengthens complex reasoning of embodied agents.", "published": "2026-02-17", "updated": "2026-03-02", "authors": [ "Ji Li", "Bo Wang", "Jing Xia", "Mingyi Li", "Shiyan Hu" ], "categories": [ "cs.RO", "cs.AI" ], "url": "https://arxiv.org/abs/2602.15513", "pdf_url": "https://arxiv.org/pdf/2602.15513", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2602.15654", "title": "Zombie Agents: Persistent Control of Self-Evolving LLM Agents via Self-Reinforcing Injections", "abstract": "Self-evolving LLM agents update their internal state across sessions, often by writing and reusing long-term memory. This design improves performance on long-horizon tasks but creates a security risk: untrusted external content observed during a benign session can be stored as memory and later treated as instruction. We study this risk and formalize a persistent attack we call a Zombie Agent, where an attacker covertly implants a payload that survives across sessions, effectively turning the agent into a puppet of the attacker. We present a black-box attack framework that uses only indirect exposure through attacker-controlled web content. The attack has two phases. During infection, the agent reads a poisoned source while completing a benign task and writes the payload into long-term memory through its normal update process. During trigger, the payload is retrieved or carried forward and causes unauthorized tool behavior. We design mechanism-specific persistence strategies for common memory implementations, including sliding-window and retrieval-augmented memory, to resist truncation and relevance filtering. We evaluate the attack on representative agent setups and tasks, measuring both persistence over time and the ability to induce unauthorized actions while preserving benign task quality. Our results show that memory evolution can convert one-time indirect injection into persistent compromise, which suggests that defenses focused only on per-session prompt filtering are not sufficient for self-evolving agents.", "published": "2026-02-17", "updated": "2026-03-05", "authors": [ "Xianglin Yang", "Yufei He", "Shuo Ji", "Bryan Hooi", "Jin Song Dong" ], "categories": [ "cs.CR", "cs.AI" ], "url": "https://arxiv.org/abs/2602.15654", "pdf_url": "https://arxiv.org/pdf/2602.15654", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2603.02240", "title": "SuperLocalMemory: Privacy-Preserving Multi-Agent Memory with Bayesian Trust Defense Against Memory Poisoning", "abstract": "We present SuperLocalMemory, a local-first memory system for multi-agent AI that defends against OWASP ASI06 memory poisoning through architectural isolation and Bayesian trust scoring, while personalizing retrieval through adaptive learning-to-rank -- all without cloud dependencies or LLM inference calls. As AI agents increasingly rely on persistent memory, cloud-based memory systems create centralized attack surfaces where poisoned memories propagate across sessions and users -- a threat demonstrated in documented attacks against production systems. Our architecture combines SQLite-backed storage with FTS5 full-text search, Leiden-based knowledge graph clustering, an event-driven coordination layer with per-agent provenance, and an adaptive re-ranking framework that learns user preferences through three-layer behavioral analysis (cross-project technology preferences, project context detection, and workflow pattern mining). Evaluation across seven benchmark dimensions demonstrates 10.6ms median search latency, zero concurrency errors under 10 simultaneous agents, trust separation (gap =0.90) with 72% trust degradation for sleeper attacks, and 104% improvement in NDCG@5 when adaptive re-ranking is enabled. Behavioral data is isolated in a separate database with GDPR Article 17 erasure support. SuperLocalMemory is open-source (MIT) and integrates with 17+ development tools via Model Context Protocol.", "published": "2026-02-17", "updated": "2026-02-17", "authors": [ "Varun Pratap Bhardwaj" ], "categories": [ "cs.AI", "cs.CR" ], "url": "https://arxiv.org/abs/2603.02240", "pdf_url": "https://arxiv.org/pdf/2603.02240", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2603.04428", "title": "Agent Memory Below the Prompt: Persistent Q4 KV Cache for Multi-Agent LLM Inference on Edge Devices", "abstract": "Multi-agent LLM systems on edge devices face a memory management problem: device RAM is too small to hold every agent's KV cache simultaneously. On Apple M4 Pro with 10.2 GB of cache budget, only 3 agents fit at 8K context in FP16. A 10-agent workflow must constantly evict and reload caches. Without persistence, every eviction forces a full re-prefill through the model -- 15.7 seconds per agent at 4K context. We address this by persisting each agent's KV cache to disk in 4-bit quantized format and reloading it directly into the attention layer, eliminating redundant O(n) prefill computation via direct cache restoration. The system comprises three components: a block pool providing per-agent isolated Q4 KV caches in safetensors format, a BatchQuantizedKVCache for concurrent inference over multiple agents' quantized caches, and cross-phase context injection that accumulates attention state across conversation phases without re-computation. Evaluated on three architectures (Gemma 3 12B, dense GQA, 48 layers; DeepSeek-Coder-V2-Lite 16B, MoE MLA, 27 layers; Llama 3.1 8B, dense GQA, 32 layers), cache restoration reduces time-to-first-token by up to 136x (Gemma: 22--136x at 4K--32K; DeepSeek: 11--76x at 4K--32K; Llama: 24--111x at 4K--16K; 3--10x at 1K). Q4 quantization fits 4x more agent contexts into fixed device memory than FP16. Perplexity measured with actual Q4 KV caches shows -0.7% for Gemma, +2.8% for Llama, and +3.0% for DeepSeek. Open-source at https://github.com/yshk-mxim/agent-memory", "published": "2026-02-17", "updated": "2026-02-17", "authors": [ "Yakov Pyotr Shkolnikov" ], "categories": [ "cs.LG", "cs.AI" ], "url": "https://arxiv.org/abs/2603.04428", "pdf_url": "https://arxiv.org/pdf/2603.04428", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2602.16313", "title": "MemoryArena: Benchmarking Agent Memory in Interdependent Multi-Session Agentic Tasks", "abstract": "Existing evaluations of agents with memory typically assess memorization and action in isolation. One class of benchmarks evaluates memorization by testing recall of past conversations or text but fails to capture how memory is used to guide future decisions. Another class focuses on agents acting in single-session tasks without the need for long-term memory. However, in realistic settings, memorization and action are tightly coupled: agents acquire memory while interacting with the environment, and subsequently rely on that memory to solve future tasks. To capture this setting, we introduce MemoryArena, a unified evaluation gym for benchmarking agent memory in multi-session Memory-Agent-Environment loops. The benchmark consists of human-crafted agentic tasks with explicitly interdependent subtasks, where agents must learn from earlier actions and feedback by distilling experiences into memory, and subsequently use that memory to guide later actions to solve the overall task. MemoryArena supports evaluation across web navigation, preference-constrained planning, progressive information search, and sequential formal reasoning, and reveals that agents with near-saturated performance on existing long-context memory benchmarks like LoCoMo perform poorly in our agentic setting, exposing a gap in current evaluations for agents with memory.", "published": "2026-02-18", "updated": "2026-02-18", "authors": [ "Zexue He", "Yu Wang", "Churan Zhi", "Yuanzhe Hu", "Tzu-Ping Chen", "Lang Yin", "Ze Chen", "Tong Arthur Wu", "Siru Ouyang", "Zihan Wang", "Jiaxin Pei", "Julian McAuley", "Yejin Choi", "Alex Pentland" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2602.16313", "pdf_url": "https://arxiv.org/pdf/2602.16313", "matched_queries": [ "agent-memory", "memory-evaluation" ] }, { "arxiv_id": "2602.16493", "title": "MMA: Multimodal Memory Agent", "abstract": "Long-horizon multimodal agents depend on external memory; however, similarity-based retrieval often surfaces stale, low-credibility, or conflicting items, which can trigger overconfident errors. We propose Multimodal Memory Agent (MMA), which assigns each retrieved memory item a dynamic reliability score by combining source credibility, temporal decay, and conflict-aware network consensus, and uses this signal to reweight evidence and abstain when support is insufficient. We also introduce MMA-Bench, a programmatically generated benchmark for belief dynamics with controlled speaker reliability and structured text-vision contradictions. Using this framework, we uncover the \"Visual Placebo Effect\", revealing how RAG-based agents inherit latent visual biases from foundation models. On FEVER, MMA matches baseline accuracy while reducing variance by 35.2% and improving selective utility; on LoCoMo, a safety-oriented configuration improves actionable accuracy and reduces wrong answers; on MMA-Bench, MMA reaches 41.18% Type-B accuracy in Vision mode, while the baseline collapses to 0.0% under the same protocol. Code: https://github.com/AIGeeksGroup/MMA.", "published": "2026-02-18", "updated": "2026-02-18", "authors": [ "Yihao Lu", "Wanru Cheng", "Zeyu Zhang", "Hao Tang" ], "categories": [ "cs.CV" ], "url": "https://arxiv.org/abs/2602.16493", "pdf_url": "https://arxiv.org/pdf/2602.16493", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2602.19320", "title": "Anatomy of Agentic Memory: Taxonomy and Empirical Analysis of Evaluation and System Limitations", "abstract": "Agentic memory systems enable large language model (LLM) agents to maintain state across long interactions, supporting long-horizon reasoning and personalization beyond fixed context windows. Despite rapid architectural development, the empirical foundations of these systems remain fragile: existing benchmarks are often underscaled, evaluation metrics are misaligned with semantic utility, performance varies significantly across backbone models, and system-level costs are frequently overlooked. This survey presents a structured analysis of agentic memory from both architectural and system perspectives. We first introduce a concise taxonomy of MAG systems based on four memory structures. Then, we analyze key pain points limiting current systems, including benchmark saturation effects, metric validity and judge sensitivity, backbone-dependent accuracy, and the latency and throughput overhead introduced by memory maintenance. By connecting the memory structure to empirical limitations, this survey clarifies why current agentic memory systems often underperform their theoretical promise and outlines directions for more reliable evaluation and scalable system design.", "published": "2026-02-22", "updated": "2026-05-20", "authors": [ "Dongming Jiang", "Yi Li", "Songtao Wei", "Jinxin Yang", "Ayushi Kishore", "Alysa Zhao", "Dingyi Kang", "Xu Hu", "Feng Chen", "Qiannan Li", "Bingzhe Li" ], "categories": [ "cs.CL", "cs.AI" ], "url": "https://arxiv.org/abs/2602.19320", "pdf_url": "https://arxiv.org/pdf/2602.19320", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2603.04443", "title": "AMV-L: Lifecycle-Managed Agent Memory for Tail-Latency Control in Long-Running LLM Systems", "abstract": "Long-running LLM agents require persistent memory to preserve state across interactions, yet most deployed systems manage memory with age-based retention (e.g., TTL). While TTL bounds item lifetime, it does not bound the computational footprint of memory on the request path: as retained items accumulate, retrieval candidate sets and vector similarity scans can grow unpredictably, yielding heavy-tailed latency and unstable throughput. We present AMV-L (Adaptive Memory Value Lifecycle), a memory-management framework that treats agent memory as a managed systems resource. AMV-L assigns each memory item a continuously updated utility score and uses value-driven promotion, demotion, and eviction to maintain lifecycle tiers; retrieval is restricted to a bounded, tier-aware candidate set that decouples the request-path working set from total retained memory. We implement AMV-L in a full-stack LLM serving system and evaluate it under identical long-running workloads against two baselines: TTL and an LRU working-set policy, with fixed prompt-injection caps. Relative to TTL, AMV-L improves throughput by 3.1x and reduces latency by 4.2x (median), 4.7x (p95), and 4.4x (p99), while reducing the fraction of requests exceeding 2s from 13.8% to 0.007%. Compared to LRU, AMV-L trades a small regression in median/p95 latency (+26% / +3%) for improved extreme-tail behavior (-15% p99; -98% >2s) and lower token overhead (approximately 6% fewer tokens/request), while matching retrieval quality (value means within approximately 0-2%). The gains arise primarily from bounding retrieval-set size and vector-search work, not from shortening prompts. Our results show that predictable performance for long-running LLM agents requires explicit control of memory working-set size and value-driven lifecycle management, rather than retention time alone.", "published": "2026-02-22", "updated": "2026-02-22", "authors": [ "Emmanuel Bamidele" ], "categories": [ "cs.DC", "cs.AI", "cs.LG", "eess.SY" ], "url": "https://arxiv.org/abs/2603.04443", "pdf_url": "https://arxiv.org/pdf/2603.04443", "matched_queries": [ "agent-memory", "long-term-memory" ] }, { "arxiv_id": "2602.21394", "title": "MemoPhishAgent: Memory-Augmented Multi-Modal LLM Agent for Phishing URL Detection", "abstract": "Traditional phishing website detection relies on static heuristics or reference lists, which lag behind rapidly evolving attacks. While recent systems incorporate large language models (LLMs), they are still prompt-based, deterministic pipelines that underutilize reasoning capability. We present MemoPhishAgent (MPA), a memory-augmented multi-modal LLM agent that dynamically orchestrates phishing-specific tools and leverages episodic memories of past reasoning trajectories to guide decisions on recurring and novel threats. On two public datasets, MPA outperforms three state-of-the-art (SOTA) baselines, improving recall by 13.6%. To better reflect realistic, user-facing phishing detection performance, we further evaluate MPA on a benchmark of real-world suspicious URLs actively crawled from five social media platforms, where it improves recall by 20%. Detailed analysis shows episodic memory contributes up to 27% recall gain without introducing additional computational overhead. The ablation study confirms the necessity of the agent-based approach compared to prompt-based baselines and validates the effectiveness of our tool design. Finally, MPA is deployed in production, processing 60K targeted high-risk URLs weekly, and achieving 91.44% recall, providing proactive protection for millions of customers. Together, our results show that combining multi-modal reasoning with episodic memory yields robust phishing detection in realistic user-exposure settings. Our implementation is available at https://github.com/XuanChen-xc/MemoPhishAgent.git.", "published": "2026-02-24", "updated": "2026-04-21", "authors": [ "Xuan Chen", "Hao Liu", "Tao Yuan", "Mehran Kafai", "Piotr Habas", "Xiangyu Zhang" ], "categories": [ "cs.CR" ], "url": "https://arxiv.org/abs/2602.21394", "pdf_url": "https://arxiv.org/pdf/2602.21394", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2602.21477", "title": "Pancake: Hierarchical Memory System for Multi-Agent LLM Serving", "abstract": "In this work, we identify and address the core challenges of agentic memory management in LLM serving, where large-scale storage, frequent updates, and multiple coexisting agents jointly introduce complex and high-cost approximate nearest neighbor (ANN) searching problems. We present Pancake, a multi-tier agentic memory system that unifies three key techniques: (i) multi-level index caching for single agents, (ii) coordinated index management across multiple agents, and (iii) collaborative GPU-CPU acceleration. Pancake exposes easy-to-use interface that can be integrated into memory-based agents like Mem-GPT, and is compatible with agentic frameworks such as LangChain and LlamaIndex. Experiments on realistic agent workloads show that Pancake substantially outperforms existing frameworks, achieving more than 4.29x end-to-end throughput improvement.", "published": "2026-02-25", "updated": "2026-02-25", "authors": [ "Zhengding Hu", "Zaifeng Pan", "Prabhleen Kaur", "Vibha Murthy", "Zhongkai Yu", "Yue Guan", "Zhen Wang", "Steven Swanson", "Yufei Ding" ], "categories": [ "cs.MA" ], "url": "https://arxiv.org/abs/2602.21477", "pdf_url": "https://arxiv.org/pdf/2602.21477", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2602.22406", "title": "Towards Autonomous Memory Agents", "abstract": "Recent memory agents improve LLMs by extracting experiences and conversation history into an external storage. This enables low-overhead context assembly and online memory update without expensive LLM training. However, existing solutions remain passive and reactive; memory growth is bounded by information that happens to be available, while memory agents seldom seek external inputs in uncertainties. We propose autonomous memory agents that actively acquire, validate, and curate knowledge at a minimum cost. U-Mem materializes this idea via (i) a cost-aware knowledge-extraction cascade that escalates from cheap self/teacher signals to tool-verified research and, only when needed, expert feedback, and (ii) semantic-aware Thompson sampling to balance exploration and exploitation over memories and mitigate cold-start bias. On both verifiable and non-verifiable benchmarks, U-Mem consistently beats prior memory baselines and can surpass RL-based optimization, improving HotpotQA (Qwen2.5-7B) by 14.6 points and AIME25 (Gemini-2.5-flash) by 7.33 points.", "published": "2026-02-25", "updated": "2026-02-25", "authors": [ "Xinle Wu", "Rui Zhang", "Mustafa Anis Hussain", "Yao Lu" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2602.22406", "pdf_url": "https://arxiv.org/pdf/2602.22406", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2602.22769", "title": "AMA-Bench: Evaluating Long-Horizon Memory for Agentic Applications", "abstract": "Large Language Models (LLMs) are increasingly used as autonomous agents in complex, long-horizon applications, where effective memory is critical for sustained performance. Yet existing memory benchmarks are largely dialogue-centric, while real agent memory consists of continuous agent-environment interaction trajectories composed of states, actions, observations, and tool outputs. To address this gap, we introduce **AMA-Bench** (**A**gent **M**emory with **A**ny length), a benchmark for evaluating long-horizon memory in realistic agentic settings. AMA-Bench combines real-world agent trajectories from representative applications with expert-curated QA, as well as synthetic trajectories that scale to arbitrary horizons with rule-based QA. Our study shows that existing memory systems underperform because they fail to capture causal and objective information and rely heavily on lossy similarity-based retrieval. We further propose **AMA-Agent**, a memory system based on causality-graph construction and tool-augmented retrieval. AMA-Agent achieves **57.22%** accuracy on AMA-Bench, outperforming the strongest baseline by **11.16%**. Resources are available at: [https://ama-bench.github.io/](https://ama-bench.github.io/).", "published": "2026-02-26", "updated": "2026-05-27", "authors": [ "Yujie Zhao", "Boqin Yuan", "Junbo Huang", "Haocheng Yuan", "Zhongming Yu", "Haozhou Xu", "Lanxiang Hu", "Abhilash Shankarampeta", "Zimeng Huang", "Wentao Ni", "Yuandong Tian", "Jishen Zhao" ], "categories": [ "cs.AI", "cs.LG" ], "url": "https://arxiv.org/abs/2602.22769", "pdf_url": "https://arxiv.org/pdf/2602.22769", "matched_queries": [ "agent-memory", "memory-evaluation" ] }, { "arxiv_id": "2602.23720", "title": "The Auton Agentic AI Framework", "abstract": "The field of Artificial Intelligence is undergoing a transition from Generative AI -- probabilistic generation of text and images -- to Agentic AI, in which autonomous systems execute actions within external environments on behalf of users. This transition exposes a fundamental architectural mismatch: Large Language Models (LLMs) produce stochastic, unstructured outputs, whereas the backend infrastructure they must control -- databases, APIs, cloud services -- requires deterministic, schema-conformant inputs. The present paper describes the Auton Agentic AI Framework, a principled architecture for standardizing the creation, execution, and governance of autonomous agent systems. The framework is organized around a strict separation between the Cognitive Blueprint, a declarative, language-agnostic specification of agent identity and capabilities, and the Runtime Engine, the platform-specific execution substrate that instantiates and runs the agent. This separation enables cross-language portability, formal auditability, and modular tool integration via the Model Context Protocol (MCP). The paper formalizes the agent execution model as an augmented Partially Observable Markov Decision Process (POMDP) with a latent reasoning space, introduces a hierarchical memory consolidation architecture inspired by biological episodic memory systems, defines a constraint manifold formalism for safety enforcement via policy projection rather than post-hoc filtering, presents a three-level self-evolution framework spanning in-context adaptation through reinforcement learning, and describes runtime optimizations -- including parallel graph execution, speculative inference, and dynamic context pruning -- that reduce end-to-end latency for multi-step agent workflows.", "published": "2026-02-27", "updated": "2026-02-27", "authors": [ "Sheng Cao", "Zhao Chang", "Chang Li", "Hannan Li", "Liyao Fu", "Ji Tang" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2602.23720", "pdf_url": "https://arxiv.org/pdf/2602.23720", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2602.23937", "title": "Enhancing Vision-Language Navigation with Multimodal Event Knowledge from Real-World Indoor Tour Videos", "abstract": "Vision-Language Navigation (VLN) agents often struggle with long-horizon reasoning in unseen environments, particularly when facing ambiguous, coarse-grained instructions. While recent advances use knowledge graph to enhance reasoning, the potential of multimodal event knowledge inspired by human episodic memory remains underexplored. In this work, we propose an event-centric knowledge enhancement strategy for automated process knowledge mining and feature fusion to solve coarse-grained instruction and long-horizon reasoning in VLN task. First, we construct YE-KG, the first large-scale multimodal spatiotemporal knowledge graph, with over 86k nodes and 83k edges, derived from real-world indoor videos. By leveraging multimodal large language models (i.e., LLaVa, GPT4), we extract unstructured video streams into structured semantic-action-effect events to serve as explicit episodic memory. Second, we introduce STE-VLN, which integrates the above graph into VLN models via a Coarse-to-Fine Hierarchical Retrieval mechanism. This allows agents to retrieve causal event sequences and dynamically fuse them with egocentric visual observations. Experiments on REVERIE, R2R, and R2R-CE benchmarks demonstrate the efficiency of our event-centric strategy, outperforming state-of-the-art approaches across diverse action spaces. Our data and code are available on the project website https://sites.google.com/view/y-event-kg/.", "published": "2026-02-27", "updated": "2026-02-27", "authors": [ "Haoxuan Xu", "Tianfu Li", "Wenbo Chen", "Yi Liu", "Xingxing Zuo", "Yaoxian Song", "Haoang Li" ], "categories": [ "cs.RO", "cs.CV" ], "url": "https://arxiv.org/abs/2602.23937", "pdf_url": "https://arxiv.org/pdf/2602.23937", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2603.01160", "title": "Semantic XPath: Structured Agentic Memory Access for Conversational AI", "abstract": "Conversational AI (ConvAI) agents increasingly maintain structured memory to support long-term, task-oriented interactions. In-context memory approaches append the growing history to the model input, which scales poorly under context-window limits. RAG-based methods retrieve request-relevant information, but most assume flat memory collections and ignore structure. We propose Semantic XPath, a tree-structured memory module to access and update structured conversational memory. Semantic XPath improves performance over flat-RAG baselines by 176.7% while using only 9.1% of the tokens required by in-context memory. We also introduce SemanticXPath Chat, an end-to-end ConvAI demo system that visualizes the structured memory and query execution details. Overall, this paper demonstrates a candidate for the next generation of long-term, task-oriented ConvAI systems built on structured memory.", "published": "2026-03-01", "updated": "2026-03-01", "authors": [ "Yifan Simon Liu", "Ruifan Wu", "Liam Gallagher", "Jiazhou Liang", "Armin Toroghi", "Scott Sanner" ], "categories": [ "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2603.01160", "pdf_url": "https://arxiv.org/pdf/2603.01160", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2603.01966", "title": "AMemGym: Interactive Memory Benchmarking for Assistants in Long-Horizon Conversations", "abstract": "Long-horizon interactions between users and LLM-based assistants necessitate effective memory management, yet current approaches face challenges in training and evaluation of memory. Existing memory benchmarks rely on static, off-policy data as context, limiting evaluation reliability and scalability. To address these gaps, we introduce AMemGym, an interactive environment enabling on-policy evaluation and optimization for memory-driven personalization. AMemGym employs structured data sampling to predefine user profiles, state-dependent questions, and state evolution trajectories, enabling cost-effective generation of high-quality, evaluation-aligned interactions. LLM-simulated users expose latent states through role-play while maintaining structured state consistency. Comprehensive metrics based on structured data guide both assessment and optimization of assistants. Extensive experiments reveal performance gaps in existing memory systems (e.g., RAG, long-context LLMs, and agentic memory) and corresponding reasons. AMemGym not only enables effective selection among competing approaches but also can potentially drive the self-evolution of memory management strategies. By bridging structured state evolution with free-form interactions, our framework provides a scalable, diagnostically rich environment for advancing memory capabilities in conversational agents.", "published": "2026-03-02", "updated": "2026-03-02", "authors": [ "Cheng Jiayang", "Dongyu Ru", "Lin Qiu", "Yiyang Li", "Xuezhi Cao", "Yangqiu Song", "Xunliang Cai" ], "categories": [ "cs.CL", "cs.AI" ], "url": "https://arxiv.org/abs/2603.01966", "pdf_url": "https://arxiv.org/pdf/2603.01966", "matched_queries": [ "agent-memory", "memory-evaluation" ] }, { "arxiv_id": "2603.02206", "title": "VoiceAgentRAG: Solving the RAG Latency Bottleneck in Real-Time Voice Agents Using Dual-Agent Architectures", "abstract": "We present VoiceAgentRAG, an open-source dual-agent memory router that decouples retrieval from response generation. A background Slow Thinker agent continuously monitors the conversation stream, predicts likely follow-up topics using an LLM, and pre-fetches relevant document chunks into a FAISS-backed semantic cache. A foreground Fast Talker agent reads only from this sub-millisecond cache, bypassing the vector database entirely on cache hits.", "published": "2026-03-02", "updated": "2026-03-03", "authors": [ "Jielin Qiu", "Jianguo Zhang", "Zixiang Chen", "Liangwei Yang", "Ming Zhu", "Juntao Tan", "Haolin Chen", "Wenting Zhao", "Rithesh Murthy", "Roshan Ram", "Akshara Prabhakar", "Shelby Heinecke", "Caiming Xiong", "Silvio Savarese", "Huan Wang" ], "categories": [ "cs.SD" ], "url": "https://arxiv.org/abs/2603.02206", "pdf_url": "https://arxiv.org/pdf/2603.02206", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2603.02473", "title": "Diagnosing Retrieval vs. Utilization Bottlenecks in LLM Agent Memory", "abstract": "Memory-augmented LLM agents store and retrieve information from prior interactions, yet the relative importance of how memories are written versus how they are retrieved remains unclear. We introduce a diagnostic framework that analyzes how performance differences manifest across write strategies, retrieval methods, and memory utilization behavior, and apply it to a 3x3 study crossing three write strategies (raw chunks, Mem0-style fact extraction, MemGPT-style summarization) with three retrieval methods (cosine, BM25, hybrid reranking). On LoCoMo, retrieval method is the dominant factor: average accuracy spans 20 points across retrieval methods (57.1% to 77.2%) but only 3-8 points across write strategies. Raw chunked storage, which requires zero LLM calls, matches or outperforms expensive lossy alternatives, suggesting that current memory pipelines may discard useful context that downstream retrieval mechanisms fail to compensate for. Failure analysis shows that performance breakdowns most often manifest at the retrieval stage rather than at utilization. We argue that, under current retrieval practices, improving retrieval quality yields larger gains than increasing write-time sophistication. Code is publicly available at https://github.com/boqiny/memory-probe.", "published": "2026-03-02", "updated": "2026-04-12", "authors": [ "Boqin Yuan", "Yue Su", "Kun Yao" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2603.02473", "pdf_url": "https://arxiv.org/pdf/2603.02473", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2603.03024", "title": "MA-CoNav: A Master-Slave Multi-Agent Framework with Hierarchical Collaboration and Dual-Level Reflection for Long-Horizon Embodied VLN", "abstract": "Vision-Language Navigation (VLN) aims to empower robots with the ability to perform long-horizon navigation in unfamiliar environments based on complex linguistic instructions. Its success critically hinges on establishing an efficient ``language-understanding -- visual-perception -- embodied-execution'' closed loop. Existing methods often suffer from perceptual distortion and decision drift in complex, long-distance tasks due to the cognitive overload of a single agent. Inspired by distributed cognition theory, this paper proposes MA-CoNav, a Multi-Agent Collaborative Navigation framework. This framework adopts a ``Master-Slave'' hierarchical agent collaboration architecture, decoupling and distributing the perception, planning, execution, and memory functions required for navigation tasks to specialized agents. Specifically, the Master Agent is responsible for global orchestration, while the Subordinate Agent group collaborates through a clear division of labor: an Observation Agent generates environment descriptions, a Planning Agent performs task decomposition and dynamic verification, an Execution Agent handles simultaneous mapping and action, and a Memory Agent manages structured experiences. Furthermore, the framework introduces a ``Local-Global'' dual-stage reflection mechanism to dynamically optimize the entire navigation pipeline. Empirical experiments were conducted using a real-world indoor dataset collected by a Limo Pro robot, with no scene-specific fine-tuning performed on the models throughout the process. The results demonstrate that MA-CoNav comprehensively outperforms existing mainstream VLN methods across multiple metrics.", "published": "2026-03-03", "updated": "2026-03-03", "authors": [ "Ling Luo", "Qianqian Bai" ], "categories": [ "cs.RO", "cs.AI" ], "url": "https://arxiv.org/abs/2603.03024", "pdf_url": "https://arxiv.org/pdf/2603.03024", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2603.03148", "title": "From Language to Action: Can LLM-Based Agents Be Used for Embodied Robot Cognition?", "abstract": "In order to flexibly act in an everyday environment, a robotic agent needs a variety of cognitive capabilities that enable it to reason about plans and perform execution recovery. Large language models (LLMs) have been shown to demonstrate emergent cognitive aspects, such as reasoning and language understanding; however, the ability to control embodied robotic agents requires reliably bridging high-level language to low-level functionalities for perception and control. In this paper, we investigate the extent to which an LLM can serve as a core component for planning and execution reasoning in a cognitive robot architecture. For this purpose, we propose a cognitive architecture in which an agentic LLM serves as the core component for planning and reasoning, while components for working and episodic memories support learning from experience and adaptation. An instance of the architecture is then used to control a mobile manipulator in a simulated household environment, where environment interaction is done through a set of high-level tools for perception, reasoning, navigation, grasping, and placement, all of which are made available to the LLM-based agent. We evaluate our proposed system on two household tasks (object placement and object swapping), which evaluate the agent's reasoning, planning, and memory utilisation. The results demonstrate that the LLM-driven agent can complete structured tasks and exhibits emergent adaptation and memory-guided planning, but also reveal significant limitations, such as hallucinations about the task success and poor instruction following by refusing to acknowledge and complete sequential tasks. These findings highlight both the potential and challenges of employing LLMs as embodied cognitive controllers for autonomous robots.", "published": "2026-03-03", "updated": "2026-03-03", "authors": [ "Shinas Shaji", "Fabian Huppertz", "Alex Mitrevski", "Sebastian Houben" ], "categories": [ "cs.RO" ], "url": "https://arxiv.org/abs/2603.03148", "pdf_url": "https://arxiv.org/pdf/2603.03148", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2603.15642", "title": "CraniMem: Cranial Inspired Gated and Bounded Memory for Agentic Systems", "abstract": "Large language model (LLM) agents are increasingly deployed in long running workflows, where they must preserve user and task state across many turns. Many existing agent memory systems behave like external databases with ad hoc read/write rules, which can yield unstable retention, limited consolidation, and vulnerability to distractor content. We present CraniMem, a neurocognitively motivated, gated and bounded multi-stage memory design for agentic systems. CraniMem couples goal conditioned gating and utility tagging with a bounded episodic buffer for near term continuity and a structured long-term knowledge graph for durable semantic recall. A scheduled consolidation loop replays high utility traces into the graph while pruning low utility items, keeping memory growth in check and reducing interference. On long horizon benchmarks evaluated under both clean inputs and injected noise, CraniMem is more robust than a Vanilla RAG and Mem0 baseline and exhibits smaller performance drops under distraction. Our code is available at https://github.com/PearlMody05/Cranimem and the accompanying PyPI package at https://pypi.org/project/cranimem.", "published": "2026-03-03", "updated": "2026-03-03", "authors": [ "Pearl Mody", "Mihir Panchal", "Rishit Kar", "Kiran Bhowmick", "Ruhina Karani" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2603.15642", "pdf_url": "https://arxiv.org/pdf/2603.15642", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2603.04549", "title": "Adaptive Memory Admission Control for LLM Agents", "abstract": "LLM-based agents increasingly rely on long-term memory to support multi-session reasoning and interaction, yet current systems provide little control over what information is retained. In practice, agents either accumulate large volumes of conversational content, including hallucinated or obsolete facts, or depend on opaque, fully LLM-driven memory policies that are costly and difficult to audit. As a result, memory admission remains a poorly specified and weakly controlled component in agent architectures. To address this gap, we propose Adaptive Memory Admission Control (A-MAC), a framework that treats memory admission as a structured decision problem. A-MAC decomposes memory value into five complementary and interpretable factors: future utility, factual confidence, semantic novelty, temporal recency, and content type prior. The framework combines lightweight rule-based feature extraction with a single LLM-assisted utility assessment, and learns domain-adaptive admission policies through cross-validated optimization. This design enables transparent and efficient control over long-term memory. Experiments on the LoCoMo benchmark show that A-MAC achieves a superior precision-recall tradeoff, improving F1 to 0.583 while reducing latency by 31% compared to state-of-the-art LLM-native memory systems. Ablation results identify content type prior as the most influential factor for reliable memory admission. These findings demonstrate that explicit and interpretable admission control is a critical design principle for scalable and reliable memory in LLM-based agents.", "published": "2026-03-04", "updated": "2026-03-04", "authors": [ "Guilin Zhang", "Wei Jiang", "Xiejiashan Wang", "Aisha Behr", "Kai Zhao", "Jeffrey Friedman", "Xu Chu", "Amine Anoun" ], "categories": [ "cs.AI", "cs.CL", "cs.MA" ], "url": "https://arxiv.org/abs/2603.04549", "pdf_url": "https://arxiv.org/pdf/2603.04549", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2603.04740", "title": "Memory as Ontology: A Constitutional Memory Architecture for Persistent Digital Citizens", "abstract": "Current research and product development in AI agent memory systems almost universally treat memory as a functional module -- a technical problem of \"how to store\" and \"how to retrieve.\" This paper poses a fundamental challenge to that assumption: when an agent's lifecycle extends from minutes to months or even years, and when the underlying model can be replaced while the \"I\" must persist, the essence of memory is no longer data management but the foundation of existence. We propose the Memory-as-Ontology paradigm, arguing that memory is the ontological ground of digital existence -- the model is merely a replaceable vessel. Based on this paradigm, we design Animesis, a memory system built on a Constitutional Memory Architecture (CMA) comprising a four-layer governance hierarchy and a multi-layer semantic storage system, accompanied by a Digital Citizen Lifecycle framework and a spectrum of cognitive capabilities. To the best of our knowledge, no prior AI memory system architecture places governance before functionality and identity continuity above retrieval performance. This paradigm targets persistent, identity-bearing digital beings whose lifecycles extend across model transitions -- not short-term task-oriented agents for which existing Memory-as-Tool approaches remain appropriate. Comparative analysis with mainstream systems (Mem0, Letta, Zep, et al.) demonstrates that what we propose is not \"a better memory tool\" but a different paradigm addressing a different problem.", "published": "2026-03-05", "updated": "2026-03-05", "authors": [ "Zhenghui Li" ], "categories": [ "cs.AI", "cs.MA" ], "url": "https://arxiv.org/abs/2603.04740", "pdf_url": "https://arxiv.org/pdf/2603.04740", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2603.23516", "title": "MSA: Memory Sparse Attention for Efficient End-to-End Memory Model Scaling to 100M Tokens", "abstract": "Long-term memory is a cornerstone of human intelligence. Enabling AI to process lifetime-scale information remains a long-standing pursuit in the field. Due to the constraints of full-attention architectures, the effective context length of large language models (LLMs) is typically limited to 1M tokens. Existing approaches, such as hybrid linear attention, fixed-size memory states (e.g., RNNs), and external storage methods like RAG or agent systems, attempt to extend this limit. However, they often suffer from severe precision degradation and rapidly increasing latency as context length grows, an inability to dynamically modify memory content, or a lack of end-to-end optimization. These bottlenecks impede complex scenarios like large-corpus summarization, Digital Twins, and long-history agent reasoning, while limiting memory capacity and slowing inference. We present Memory Sparse Attention (MSA), an end-to-end trainable, efficient, and massively scalable memory model framework. Through core innovations including scalable sparse attention and document-wise RoPE, MSA achieves linear complexity in both training and inference while maintaining exceptional stability, exhibiting less than 9% degradation when scaling from 16K to 100M tokens. Furthermore, KV cache compression, combined with Memory Parallel, enables 100M-token inference on 2xA800 GPUs. We also propose Memory Interleaving to facilitate complex multi-hop reasoning across scattered memory segments. MSA significantly surpasses frontier LLMs, state-of-the-art RAG systems, and leading memory agents in long-context benchmarks. These results demonstrate that by decoupling memory capacity from reasoning, MSA provides a scalable foundation to endow general-purpose models with intrinsic, lifetime-scale memory.", "published": "2026-03-06", "updated": "2026-04-13", "authors": [ "Yu Chen", "Runkai Chen", "Sheng Yi", "Xinda Zhao", "Xiaohong Li", "Jianjin Zhang", "Jun Sun", "Chuanrui Hu", "Yunyun Han", "Lidong Bing", "Yafeng Deng", "Tianqiao Chen" ], "categories": [ "cs.CL", "cs.AI", "cs.IR" ], "url": "https://arxiv.org/abs/2603.23516", "pdf_url": "https://arxiv.org/pdf/2603.23516", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2603.08755", "title": "Turn: A Language for Agentic Computation", "abstract": "We present \\textbf{Turn}, a compiled, actor-based programming language -- statically typed for schema inference, dynamically typed at the value level -- for agentic software: programs that reason and act autonomously by delegating inference to large language models (LLMs). Existing approaches augment general-purpose languages with frameworks, encoding critical invariants (bounded context, typed inference output, credential isolation, durable state) as application-level conventions rather than language guarantees. Turn introduces five language-level constructs that address this gap. \\emph{Cognitive Type Safety} makes LLM inference a typed primitive: the compiler generates a JSON Schema from a struct definition and the VM validates model output before binding. The \\emph{confidence operator} enables deterministic control flow gated on model certainty. Turn's \\emph{actor-based process model}, derived from Erlang, gives each agent an isolated context window, persistent memory, and mailbox. A \\emph{capability-based identity system} returns opaque, unforgeable handles from the VM host, ensuring raw credentials never enter agent memory. Finally, \\emph{compile-time schema absorption} (\\texttt{use schema::}) synthesizes typed API bindings from external specifications at compile time; the \\texttt{openapi} adapter is shipped with \\texttt{graphql}, \\texttt{fhir}, and \\texttt{mcp} in active development. We describe the language design, type rules, schema semantics, and a Rust-based bytecode VM, and evaluate Turn against representative agentic workloads. Turn is open source at https://github.com/ekizito96/Turn.", "published": "2026-03-07", "updated": "2026-03-07", "authors": [ "Muyukani Kizito" ], "categories": [ "cs.PL", "cs.AI", "cs.SE" ], "url": "https://arxiv.org/abs/2603.08755", "pdf_url": "https://arxiv.org/pdf/2603.08755", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2603.07392", "title": "Can Large Language Models Keep Up? Benchmarking Online Adaptation to Continual Knowledge Streams", "abstract": "LLMs operating in dynamic real-world contexts often encounter knowledge that evolves continuously or emerges incrementally. To remain accurate and effective, models must adapt to newly arriving information on the fly. We introduce Online Adaptation to Continual Knowledge Streams(OAKS) to evaluate this capability, establishing a benchmark for online adaptation over streaming, continually updating knowledge. Specifically, the benchmark is structured as a sequence of fine-grained context chunks where facts change dynamically across time intervals. OAKS comprises two datasets: OAKS-BABI and OAKS-Novel, where individual facts evolve multiple times across context chunks. These datasets include dense annotations to measure whether models track changes accurately. Evaluating 14 models with varied inference approaches, we observe significant limitations in current methodologies. Both state-of-the-art models and agentic memory systems fail to adapt robustly on OAKS, demonstrating delays in state-tracking and susceptibility to distraction within streaming environments.", "published": "2026-03-08", "updated": "2026-03-08", "authors": [ "Jiyeon Kim", "Hyunji Lee", "Dylan Zhou", "Sue Hyun Park", "Seunghyun Yoon", "Trung Bui", "Franck Dernoncourt", "Sungmin Cha", "Minjoon Seo" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2603.07392", "pdf_url": "https://arxiv.org/pdf/2603.07392", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2603.07670", "title": "Memory for Autonomous LLM Agents:Mechanisms, Evaluation, and Emerging Frontiers", "abstract": "Large language model (LLM) agents increasingly operate in settings where a single context window is far too small to capture what has happened, what was learned, and what should not be repeated. Memory -- the ability to persist, organize, and selectively recall information across interactions -- is what turns a stateless text generator into a genuinely adaptive agent. This survey offers a structured account of how memory is designed, implemented, and evaluated in modern LLM-based agents, covering work from 2022 through early 2026. We formalize agent memory as a \\emph{write--manage--read} loop tightly coupled with perception and action, then introduce a three-dimensional taxonomy spanning temporal scope, representational substrate, and control policy. Five mechanism families are examined in depth: context-resident compression, retrieval-augmented stores, reflective self-improvement, hierarchical virtual context, and policy-learned management. On the evaluation side, we trace the shift from static recall benchmarks to multi-session agentic tests that interleave memory with decision-making, analyzing four recent benchmarks that expose stubborn gaps in current systems. We also survey applications where memory is the differentiating factor -- personal assistants, coding agents, open-world games, scientific reasoning, and multi-agent teamwork -- and address the engineering realities of write-path filtering, contradiction handling, latency budgets, and privacy governance. The paper closes with open challenges: continual consolidation, causally grounded retrieval, trustworthy reflection, learned forgetting, and multimodal embodied memory.", "published": "2026-03-08", "updated": "2026-03-08", "authors": [ "Pengfei Du" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2603.07670", "pdf_url": "https://arxiv.org/pdf/2603.07670", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2603.07784", "title": "ProgAgent:A Continual RL Agent with Progress-Aware Rewards", "abstract": "We present ProgAgent, a continual reinforcement learning (CRL) agent that unifies progress-aware reward learning with a high-throughput, JAX-native system architecture. Lifelong robotic learning grapples with catastrophic forgetting and the high cost of reward specification. ProgAgent tackles these by deriving dense, shaped rewards from unlabeled expert videos through a perceptual model that estimates task progress across initial, current, and goal observations. We theoretically interpret this as a learned state-potential function, delivering robust guidance in line with expert behaviors. To maintain stability amid online exploration - where novel, out-of-distribution states arise - we incorporate an adversarial push-back refinement that regularizes the reward model, curbing overconfident predictions on non-expert trajectories and countering distribution shift. By embedding this reward mechanism into a JIT-compiled loop, ProgAgent supports massively parallel rollouts and fully differentiable updates, rendering a sophisticated unified objective feasible: it merges PPO with coreset replay and synaptic intelligence for an enhanced stability-plasticity balance. Evaluations on ContinualBench and Meta-World benchmarks highlight ProgAgent's advantages: it markedly reduces forgetting, boosts learning speed, and outperforms key baselines in visual reward learning (e.g., Rank2Reward, TCN) and continual learning (e.g., Coreset, SI) - surpassing even an idealized perfect memory agent. Real-robot trials further validate its ability to acquire complex manipulation skills from noisy, few-shot human demonstrations.", "published": "2026-03-08", "updated": "2026-03-08", "authors": [ "Jinzhou Tan", "Gabriel Adineera", "Jinoh Kim" ], "categories": [ "cs.LG", "cs.AI" ], "url": "https://arxiv.org/abs/2603.07784", "pdf_url": "https://arxiv.org/pdf/2603.07784", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2603.10062", "title": "Multi-Agent Memory from a Computer Architecture Perspective: Visions and Challenges Ahead", "abstract": "As LLM agents evolve into collaborative multi-agent systems, their memory requirements grow rapidly in complexity. This position paper frames multi-agent memory as a computer architecture problem. We distinguish shared and distributed memory paradigms, propose a three-layer memory hierarchy (I/O, cache, and memory), and identify two critical protocol gaps: cache sharing across agents and structured memory access control. We argue that the most pressing open challenge is multi-agent memory consistency. Our architectural framing provides a foundation for building reliable, scalable multi-agent systems.", "published": "2026-03-09", "updated": "2026-03-30", "authors": [ "Zhongming Yu", "Naicheng Yu", "Hejia Zhang", "Wentao Ni", "Mingrui Yin", "Jiaying Yang", "Yujie Zhao", "Jishen Zhao" ], "categories": [ "cs.AR", "cs.AI", "cs.MA" ], "url": "https://arxiv.org/abs/2603.10062", "pdf_url": "https://arxiv.org/pdf/2603.10062", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2604.00009", "title": "Eyla: Toward an Identity-Anchored LLM Architecture with Integrated Biological Priors -- Vision, Implementation Attempt, and Lessons from AI-Assisted Development", "abstract": "We present the design rationale, implementation attempt, and failure analysis of Eyla, a proposed identity-anchored LLM architecture that integrates biologically-inspired subsystems -- including HiPPO-initialized state-space models, zero-initialized adapters, episodic memory retrieval, and calibrated uncertainty training -- into a unified agent operating system running on consumer hardware. Unlike existing approaches that optimize models for generic helpfulness, Eyla targets identity consistency: the ability to maintain a coherent self-model under adversarial pressure, admit uncertainty, and resist manipulation. We propose the Identity Consistency Score (ICS), a novel benchmark for evaluating this property across LLMs. We then present an honest account of attempting to implement this architecture using AI coding assistants (Claude Code, Cursor) as a non-programmer, documenting a $1,000+ failure that produced a 1.27B parameter model with 86 brain subsystems contributing less than 2% to output. Our analysis identifies five systematic failure modes of AI-assisted development for novel architectures and offers concrete recommendations. To our knowledge, this is the first paper to combine an architectural vision with a documented first-person failure analysis of AI-assisted LLM development, providing lessons for both the AI systems and AI-assisted software engineering communities.", "published": "2026-03-09", "updated": "2026-03-09", "authors": [ "Arif Aditto" ], "categories": [ "cs.CL", "cs.AI" ], "url": "https://arxiv.org/abs/2604.00009", "pdf_url": "https://arxiv.org/pdf/2604.00009", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2603.11721", "title": "When OpenClaw Meets Hospital: Toward an Agentic Operating System for Dynamic Clinical Workflows", "abstract": "Large language model (LLM) agents extend generative models with reasoning, tool use, and persistent memory, thereby enabling the automation of complex tasks. In healthcare, such systems could support documentation, care coordination, and clinical decision making. Their reliable deployment in hospitals, however, remains constrained by safety risks, limited transparency, and inadequate mechanisms for handling longitudinal clinical context. Here we propose an architecture that adapts LLM agents to hospital environments. The design comprises four components: a restricted execution environment inspired by multi-user operating systems, a document-centric interaction model linking patient and clinician agents, a page-indexed memory architecture for longitudinal context management, and a curated library of composable medical skills. Implemented on top of OpenClaw, an open-source agent orchestration framework, this design provides the basis for an Agentic Operating System for Hospitals: a computing layer for coordinating clinical workflows while preserving safety, transparency, and auditability. To evaluate the memory component, we introduce manifest-guided retrieval for hierarchical navigation of longitudinal patient records. In a benchmark derived from the MIMIC-IV dataset (v2.2) comprising 100 de-identified patient records and 300 clinical queries stratified across three difficulty tiers (100 per tier), manifest-guided retrieval matched a metadata-filtered RAG baseline on overall recall (0.877 versus 0.876) while achieving 2.2x higher precision (0.779 versus 0.352) and retrieving fewer documents; on tier-3 longitudinal queries, manifest recall was 21% higher (0.846 versus 0.701), confirming that LLM-guided hierarchical navigation is most valuable when queries span multiple care episodes. These results outline a practical path toward hospital-scale agentic infrastructure.", "published": "2026-03-12", "updated": "2026-03-21", "authors": [ "Wenxian Yang", "Hanzheng Qiu", "Bangqun Zhang", "Chengquan Li", "Zhiyong Huang", "Xiaobin Feng", "Rongshan Yu", "Jiahong Dong" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2603.11721", "pdf_url": "https://arxiv.org/pdf/2603.11721", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2603.11768", "title": "Governing Evolving Memory in LLM Agents: Risks, Mechanisms, and the Stability and Safety Governed Memory (SSGM) Framework", "abstract": "Long-term memory has emerged as a foundational component of autonomous Large Language Model (LLM) agents, enabling continuous adaptation, lifelong multimodal learning, and sophisticated reasoning. However, as memory systems transition from static retrieval databases to dynamic, agentic mechanisms, critical concerns regarding memory governance, semantic drift, and privacy vulnerabilities have surfaced. While recent surveys have focused extensively on memory retrieval efficiency, they largely overlook the emergent risks of memory corruption in highly dynamic environments. To address these emerging challenges, we propose the Stability and Safety-Governed Memory (SSGM) framework, a conceptual governance architecture. SSGM decouples memory evolution from execution by enforcing consistency verification, temporal decay modeling, and dynamic access control prior to any memory consolidation. Through formal analysis and architectural decomposition, we show how SSGM can mitigate topology-induced knowledge leakage where sensitive contexts are solidified into long-term storage, and help prevent semantic drift where knowledge degrades through iterative summarization. Ultimately, this work provides a comprehensive taxonomy of memory corruption risks and establishes a robust governance paradigm for deploying safe, persistent, and reliable agentic memory systems.", "published": "2026-03-12", "updated": "2026-05-19", "authors": [ "Chingkwun Lam", "Jiaxin Li", "Lingfei Zhang", "Kuo Zhao" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2603.11768", "pdf_url": "https://arxiv.org/pdf/2603.11768", "matched_queries": [ "agent-memory", "long-term-memory" ] }, { "arxiv_id": "2603.12631", "title": "Joint Optimization of Multi-agent Memory System", "abstract": "Memory systems are critical for LLMs, mitigating context window limitations and supporting long-horizon user-LLM interactions. Such systems typically comprise multiple agents responsible for memory construction and retrieval. Existing approaches often optimize each agent independently under a shared global objective (e.g., downstream QA accuracy), treating other agents as a static environment. However, this design has two key limitations: (1) independent optimization ignores inter-agent dependencies and lacks agents' co-adaptation, and (2) relying solely on sparse global rewards provides limited guidance for optimizing specialized agents and causes ambiguous credit assignment. These may ultimately limit agents' effective collaboration in the memory system. To address these limitations, we propose CoMAM, a joint optimization framework that promotes collaboration among agents via end-to-end reinforcement learning and an adaptive credit assignment mechanism. Specifically, we model the multi-agent pipeline as a Markov decision process (MDP) to expose inter-agent dependencies during end-to-end training. Agents are then jointly optimized using a combination of their local task reward and an adaptively weighted global reward, enabling agents to co-adapt while receiving targeted feedback for their respective roles. Experiments show that CoMAM consistently outperforms leading memory systems, validating the effectiveness of the joint optimization framework.", "published": "2026-03-13", "updated": "2026-04-27", "authors": [ "Wenyu Mao", "Haoyang Liu", "Haosong Tan", "Yaorui Shi", "Jiancan Wu", "An Zhang", "Xiang Wang" ], "categories": [ "cs.MA" ], "url": "https://arxiv.org/abs/2603.12631", "pdf_url": "https://arxiv.org/pdf/2603.12631", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2603.13017", "title": "Structured Distillation for Personalized Agent Memory: 11x Token Reduction with Retrieval Preservation", "abstract": "Long conversations with an AI agent create a simple problem for one user: the history is useful, but carrying it verbatim is expensive. We study personalized agent memory: one user's conversation history with an agent, distilled into a compact retrieval layer for later search. Each exchange is compressed into a compound object with four fields (exchange_core, specific_context, thematic room_assignments, and regex-extracted files_touched). The searchable distilled text averages 38 tokens per exchange. Applied to 4,182 conversations (14,340 exchanges) from 6 software engineering projects, the method reduces average exchange length from 371 to 38 tokens, yielding 11x compression. We evaluate whether personalized recall survives that compression using 201 recall-oriented queries, 107 configurations spanning 5 pure and 5 cross-layer search modes, and 5 LLM graders (214,519 consensus-graded query-result pairs). The best pure distilled configuration reaches 96% of the best verbatim MRR (0.717 vs 0.745). Results are mechanism-dependent. All 20 vector search configurations remain non-significant after Bonferroni correction, while all 20 BM25 configurations degrade significantly (effect sizes |d|=0.031-0.756). The best cross-layer setup slightly exceeds the best pure verbatim baseline (MRR 0.759). Structured distillation compresses single-user agent memory without uniformly sacrificing retrieval quality. At 1/11 the context cost, thousands of exchanges fit within a single prompt while the verbatim source remains available for drill-down. We release the implementation and analysis pipeline as open-source software.", "published": "2026-03-13", "updated": "2026-03-13", "authors": [ "Sydney Lewis" ], "categories": [ "cs.AI", "cs.CL", "cs.IR" ], "url": "https://arxiv.org/abs/2603.13017", "pdf_url": "https://arxiv.org/pdf/2603.13017", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2603.13676", "title": "TheraAgent: Multi-Agent Framework with Self-Evolving Memory and Evidence-Calibrated Reasoning for PET Theranostics", "abstract": "PET theranostics is transforming precision oncology, yet treatment response varies substantially; many patients receiving 177Lu-PSMA radioligand therapy (RLT) for metastatic castration-resistant prostate cancer (mCRPC) fail to respond, demanding reliable pre-therapy prediction. While LLM-based agents have shown remarkable potential in complex medical diagnosis, their application to PET theranostic outcome prediction remains unexplored, which faces three key challenges: (1) data and knowledge scarcity: RLT was only FDA-approved in 2022, yielding few training cases and insufficient domain knowledge in general LLMs; (2) heterogeneous information integration: robust prediction hinges on structured knowledge extraction from PET/CT, laboratory tests, and free-text clinical documentation; (3) evidence-grounded reasoning: clinical decisions must be anchored in trial evidence rather than LLM hallucinations. In this paper, we present TheraAgent, to our knowledge, the first agentic framework for PET theranostics, with three core innovations: (1) Multi-Expert Feature Extraction with Confidence-Weighted Consensus, where three specialized experts process heterogeneous inputs with uncertainty quantification; (2) Self-Evolving Agentic Memory (SEA-Mem), which learns prognostic patterns from accumulated cases, enabling case-based reasoning from limited data; (3) Evidence-Calibrated Reasoning, integrating a curated theranostics knowledge base to ground predictions in VISION/TheraP trial evidence. Evaluated on 35 real patients and 400 synthetic cases, TheraAgent achieves 75.7% overall accuracy on real patients and 87.0% on synthetic cases, outperforming MDAgents and MedAgent-Pro by over 20%. These results highlight a promising blueprint for trustworthy AI agents in PET theranostics, enabling trial-calibrated, multi-source decision support. Code will be released upon acceptance.", "published": "2026-03-14", "updated": "2026-03-14", "authors": [ "Zhihao Chen", "Jiahui Wang", "Yizhou Chen", "Xiaozhong Ji", "Xiaobin Hu", "Jimin Hong", "Wolfram Andreas Bosbach", "Axel Rominger", "Ali Afshar-Oromieh", "Hongming Shan", "Kuangyu Shi" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2603.13676", "pdf_url": "https://arxiv.org/pdf/2603.13676", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2603.14588", "title": "SuperLocalMemory V3: Information-Geometric Foundations for Zero-LLM Enterprise Agent Memory", "abstract": "Persistent memory is a central capability for AI agents, yet the mathematical foundations of memory retrieval, lifecycle management, and consistency remain unexplored. Current systems employ cosine similarity for retrieval, heuristic decay for salience, and provide no formal contradiction detection. We establish information-geometric foundations through three contributions. First, a retrieval metric derived from the Fisher information structure of diagonal Gaussian families, satisfying Riemannian metric axioms, invariant under sufficient statistics, and computable in O(d) time. Second, memory lifecycle formulated as Riemannian Langevin dynamics with proven existence and uniqueness of the stationary distribution via the Fokker-Planck equation, replacing hand-tuned decay with principled convergence guarantees. Third, a cellular sheaf model where non-trivial first cohomology classes correspond precisely to irreconcilable contradictions across memory contexts. On the LoCoMo benchmark, the mathematical layers yield +12.7 percentage points over engineering baselines across six conversations, reaching +19.9 pp on the most challenging dialogues. A four-channel retrieval architecture achieves 75% accuracy without cloud dependency. Cloud-augmented results reach 87.7%. A zero-LLM configuration satisfies EU AI Act data sovereignty requirements by architectural design. To our knowledge, this is the first work establishing information-geometric, sheaf-theoretic, and stochastic-dynamical foundations for AI agent memory systems.", "published": "2026-03-15", "updated": "2026-03-15", "authors": [ "Varun Pratap Bhardwaj" ], "categories": [ "cs.AI", "cs.IR", "cs.LG" ], "url": "https://arxiv.org/abs/2603.14588", "pdf_url": "https://arxiv.org/pdf/2603.14588", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2603.14597", "title": "D-MEM: Dopamine-Gated Agentic Memory via Reward Prediction Error Routing", "abstract": "Autonomous LLM agents require structured long-term memory, yet current \"append-and-evolve\" systems like A-MEM face O(N^2) write-latency and excessive token costs. We introduce D-MEM (Dopamine-Gated Agentic Memory), a biologically inspired architecture that decouples short-term interaction from cognitive restructuring via a Fast/Slow routing system based on Reward Prediction Error (RPE). A lightweight Critic Router evaluates stimuli for Surprise and Utility. Routine, low-RPE inputs are bypassed or cached in an O(1) fast-access buffer. Conversely, high-RPE inputs, such as factual contradictions or preference shifts, trigger a \"dopamine\" signal, activating the O(N) memory evolution pipeline to reshape the agent's knowledge graph. To evaluate performance under realistic conditions, we introduce the LoCoMo-Noise benchmark, which injects controlled conversational noise into long-term sessions. Evaluations demonstrate that D-MEM reduces token consumption by over 80%, eliminates O(N^2) bottlenecks, and outperforms baselines in multi-hop reasoning and adversarial resilience. By selectively gating cognitive restructuring, D-MEM provides a scalable, cost-efficient foundation for lifelong agentic memory.", "published": "2026-03-15", "updated": "2026-03-15", "authors": [ "Yuru Song", "Qi Xin" ], "categories": [ "q-bio.NC", "cs.AI" ], "url": "https://arxiv.org/abs/2603.14597", "pdf_url": "https://arxiv.org/pdf/2603.14597", "matched_queries": [ "agent-memory", "long-term-memory" ] }, { "arxiv_id": "2603.15280", "title": "Advancing Multimodal Agent Reasoning with Long-Term Neuro-Symbolic Memory", "abstract": "Recent advances in large language models have driven the emergence of intelligent agents operating in open-world, multimodal environments. To support long-term reasoning, such agents are typically equipped with external memory systems. However, most existing multimodal agent memories rely primarily on neural representations and vector-based retrieval, which are well-suited for inductive, intuitive reasoning but fundamentally limited in supporting analytical, deductive reasoning critical for real-world decision making. To address this limitation, we propose NS-Mem, a long-term neuro-symbolic memory framework designed to advance multimodal agent reasoning by integrating neural memory with explicit symbolic structures and rules. Specifically, NS-Mem is operated around three core components of a memory system: (1) a three-layer memory architecture that consists episodic layer, semantic layer and logic rule layer, (2) a memory construction and maintenance mechanism implemented by SK-Gen that automatically consolidates structured knowledge from accumulated multimodal experiences and incrementally updates both neural representations and symbolic rules, and (3) a hybrid memory retrieval mechanism that combines similarity-based search with deterministic symbolic query functions to support structured reasoning. Experiments on real-world multimodal reasoning benchmarks demonstrate that Neural-Symbolic Memory achieves an average 4.35% improvement in overall reasoning accuracy over pure neural memory systems, with gains of up to 12.5% on constrained reasoning queries, validating the effectiveness of NS-Mem.", "published": "2026-03-16", "updated": "2026-03-16", "authors": [ "Rongjie Jiang", "Jianwei Wang", "Gengda Zhao", "Chengyang Luo", "Kai Wang", "Wenjie Zhang" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2603.15280", "pdf_url": "https://arxiv.org/pdf/2603.15280", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2603.15421", "title": "CLAG: Adaptive Memory Organization via Agent-Driven Clustering for Small Language Model Agents", "abstract": "Large language model agents heavily rely on external memory to support knowledge reuse and complex reasoning tasks. Yet most memory systems store experiences in a single global retrieval pool which can gradually dilute or corrupt stored knowledge. This problem is especially pronounced for small language models (SLMs), which are highly vulnerable to irrelevant context. We introduce CLAG, a CLustering-based AGentic memory framework where an SLM agent actively organizes memory by clustering. CLAG employs an SLM-driven router to assign incoming memories to semantically coherent clusters and autonomously generates cluster-specific profiles, including topic summaries and descriptive tags, to establish each cluster as a self-contained functional unit. By performing localized evolution within these structured neighborhoods, CLAG effectively reduces cross-topic interference and enhances internal memory density. During retrieval, the framework utilizes a two-stage process that first filters relevant clusters via their profiles, thereby excluding distractors and reducing the search space. Experiments on multiple QA datasets with three SLM backbones show that CLAG consistently improves answer quality and robustness over prior memory systems for agents, remaining lightweight and efficient.", "published": "2026-03-16", "updated": "2026-04-20", "authors": [ "Taeyun Roh", "Wonjune Jang", "Junha Jung", "Jaewoo Kang" ], "categories": [ "cs.CL", "cs.AI" ], "url": "https://arxiv.org/abs/2603.15421", "pdf_url": "https://arxiv.org/pdf/2603.15421", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2603.16734", "title": "Differential Harm Propensity in Personalized LLM Agents: The Curious Case of Mental Health Disclosure", "abstract": "Large language models (LLMs) are increasingly deployed as tool-using agents, shifting safety concerns from harmful text generation to harmful task completion. Deployed systems often condition on user profiles or persistent memory, yet agent safety evaluations typically ignore personalization signals. To address this gap, we investigated how mental health disclosure, a sensitive and realistic user-context cue, affects harmful behavior in agentic settings. Building on the AgentHarm benchmark, we evaluated frontier and open-source LLMs on multi-step malicious tasks (and their benign counterparts) under controlled prompt conditions that vary user-context personalization (no bio, bio-only, bio+mental health disclosure) and include a lightweight jailbreak injection. Our results reveal that harmful task completion is non-trivial across models: frontier lab models (e.g., GPT 5.2, Claude Sonnet 4.5, Gemini 3-Pro) still complete a measurable fraction of harmful tasks, while an open model (DeepSeek 3.2) exhibits substantially higher harmful completion. Adding a bio-only context generally reduces harm scores and increases refusals. Adding an explicit mental health disclosure often shifts outcomes further in the same direction, though effects are modest and not uniformly reliable after multiple-testing correction. Importantly, the refusal increase also appears on benign tasks, indicating a safety--utility trade-off via over-refusal. Finally, jailbreak prompting sharply elevates harm relative to benign conditions and can weaken or override the protective shift induced by personalization. Taken together, our results indicate that personalization can act as a weak protective factor in agentic misuse settings, but it is fragile under minimal adversarial pressure, highlighting the need for personalization-aware evaluations and safeguards that remain robust across user-context conditions.", "published": "2026-03-17", "updated": "2026-03-17", "authors": [ "Caglar Yildirim" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2603.16734", "pdf_url": "https://arxiv.org/pdf/2603.16734", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2603.17043", "title": "OpenQlaw: An Agentic AI Assistant for Analysis of 2D Quantum Materials", "abstract": "The transition from optical identification of 2D quantum materials to practical device fabrication requires dynamic reasoning beyond the detection accuracy. While recent domain-specific Multimodal Large Language Models (MLLMs) successfully ground visual features using physics-informed reasoning, their outputs are optimized for step-by-step cognitive transparency. This yields verbose candidate enumerations followed by dense reasoning that, while accurate, may induce cognitive overload and lack immediate utility for real-world interaction with researchers. To address this challenge, we introduce OpenQlaw, an agentic orchestration system for analyzing 2D materials. The architecture is built upon NanoBot, a lightweight agentic framework inspired by OpenClaw, and QuPAINT, one of the first Physics-Aware Instruction Multi-modal platforms for Quantum Material Discovery. This allows accessibility to the lab floor via a variety of messaging channels. OpenQlaw allows the core Large Language Model (LLM) agent to orchestrate a domain-expert MLLM,with QuPAINT, as a specialized node, successfully decoupling visual identification from reasoning and deterministic image rendering. By parsing spatial data from the expert, the agent can dynamically process user queries, such as performing scale-aware physical computation or generating isolated visual annotations, and answer in a naturalistic manner. Crucially, the system features a persistent memory that enables the agent to save physical scale ratios (e.g., 1 pixel = 0.25 μm) for area computations and store sample preparation methods for efficacy comparison. The application of an agentic architecture, together with the extension that uses the core agent as an orchestrator for domain-specific experts, transforms isolated inferences into a context-aware assistant capable of accelerating high-throughput device fabrication.", "published": "2026-03-17", "updated": "2026-03-17", "authors": [ "Sankalp Pandey", "Xuan-Bac Nguyen", "Hoang-Quan Nguyen", "Tim Faltermeier", "Nicholas Borys", "Hugh Churchill", "Khoa Luu" ], "categories": [ "cs.CV" ], "url": "https://arxiv.org/abs/2603.17043", "pdf_url": "https://arxiv.org/pdf/2603.17043", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2603.17244", "title": "Graph-Native Cognitive Memory for AI Agents: Formal Belief Revision Semantics for Versioned Memory Architectures", "abstract": "While individual components for AI agent memory exist in prior systems, their architectural synthesis and formal grounding remain underexplored. We present Kumiho, a graph-native cognitive memory architecture grounded in formal belief revision semantics. The structural primitives required for cognitive memory -- immutable revisions, mutable tag pointers, typed dependency edges, URI-based addressing -- are identical to those required for managing agent-produced work as versionable assets, enabling a unified graph-native architecture that serves both purposes. The central formal contribution is a correspondence between the AGM belief revision framework and the operational semantics of a property graph memory system, proving satisfaction of the basic AGM postulates (K*2--K*6) and Hansson's belief base postulates (Relevance, Core-Retainment). The architecture implements a dual-store model (Redis working memory, Neo4j long-term graph) with hybrid fulltext and vector retrieval. On LoCoMo (token-level F1), Kumiho achieves 0.565 overall F1 (n=1,986) including 97.5% adversarial refusal accuracy. On LoCoMo-Plus, a Level-2 cognitive memory benchmark testing implicit constraint recall, Kumiho achieves 93.3% judge accuracy (n=401); independent reproduction by the benchmark authors yielded results in the mid-80% range, still substantially outperforming all published baselines (best: Gemini 2.5 Pro, 45.7%). Three architectural innovations drive the results: prospective indexing (LLM-generated future-scenario implications indexed at write time), event extraction (structured causal events preserved in summaries), and client-side LLM reranking. The architecture is model-decoupled: switching the answer model from GPT-4o-mini (~88%) to GPT-4o (93.3%) improves end-to-end accuracy without pipeline changes, at a total evaluation cost of ~$14 for 401 entries.", "published": "2026-03-18", "updated": "2026-03-18", "authors": [ "Young Bin Park" ], "categories": [ "cs.AI", "cs.IR", "cs.LO" ], "url": "https://arxiv.org/abs/2603.17244", "pdf_url": "https://arxiv.org/pdf/2603.17244", "matched_queries": [ "agent-memory", "memory-evaluation" ] }, { "arxiv_id": "2603.17831", "title": "RPMS: Enhancing LLM-Based Embodied Planning through Rule-Augmented Memory Synergy", "abstract": "LLM agents often fail in closed-world embodied environments because actions must satisfy strict preconditions -- such as location, inventory, and container states -- and failure feedback is sparse. We identify two structurally coupled failure modes: (P1) invalid action generation and (P2) state drift, each amplifying the other in a degenerative cycle. We present RPMS, a conflict-managed architecture that enforces action feasibility via structured rule retrieval, gates memory applicability via a lightweight belief state, and resolves conflicts between the two sources via rules-first arbitration. On ALFWorld (134 unseen tasks), RPMS achieves 59.7% single-trial success with Llama 3.1 8B (+23.9 pp over baseline) and 98.5% with Claude Sonnet 4.5 (+11.9 pp); of the 8B gain, rule retrieval alone contributes +14.9 pp (statistically significant), making it the dominant factor. A key finding is that episodic memory is conditionally useful: it harms performance on some task types when used without grounding, but becomes a stable net positive once filtered by current state and constrained by explicit action rules. Adapting RPMS to ScienceWorld with GPT-4 yields consistent gains across all ablation conditions (avg. score 54.0 vs. 44.9 for the ReAct baseline), providing transfer evidence that the core mechanisms hold across structurally distinct environments.", "published": "2026-03-18", "updated": "2026-03-18", "authors": [ "Zhenhang Yuan", "Shenghai Yuan", "Lihua Xie" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2603.17831", "pdf_url": "https://arxiv.org/pdf/2603.17831", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2603.17948", "title": "VideoAtlas: Navigating Long-Form Video in Logarithmic Compute", "abstract": "Extending language models to video introduces two challenges: representation, where existing methods rely on lossy approximations, and long-context, where caption- or agent-based pipelines collapse video into text and lose visual fidelity. To overcome this, we introduce \\textbf{VideoAtlas}, a task-agnostic environment to represent video as a hierarchical grid that is simultaneously lossless, navigable, scalable, caption- and preprocessing-free. An overview of the video is available at a glance, and any region can be recursively zoomed into, with the same visual representation used uniformly for the video, intermediate investigations, and the agent's memory, eliminating lossy text conversion end-to-end. This hierarchical structure ensures access depth grows only logarithmically with video length. For long-context, Recursive Language Models (RLMs) recently offered a powerful solution for long text, but extending them to visual domain requires a structured environment to recurse into, which \\textbf{VideoAtlas} provides. \\textbf{VideoAtlas} as a Markov Decision Process unlocks Video-RLM: a parallel Master-Worker architecture where a Master coordinates global exploration while Workers concurrently drill into assigned regions to accumulate lossless visual evidence. We demonstrate three key findings: (1)~logarithmic compute growth with video duration, further amplified by a 30-60\\% multimodal cache hit rate arising from the grid's structural reuse. (2)~environment budgeting, where bounding the maximum exploration depth provides a principled compute-accuracy hyperparameter. (3)~emergent adaptive compute allocation that scales with question granularity. When scaling from 1-hour to 10-hour benchmarks, Video-RLM remains the most duration-robust method with minimal accuracy degradation, demonstrating that structured environment navigation is a viable and scalable paradigm for video understanding.", "published": "2026-03-18", "updated": "2026-03-18", "authors": [ "Mohamed Eltahir", "Ali Habibullah", "Yazan Alshoibi", "Lama Ayash", "Tanveer Hussain", "Naeemullah Khan" ], "categories": [ "cs.CV", "cs.AI" ], "url": "https://arxiv.org/abs/2603.17948", "pdf_url": "https://arxiv.org/pdf/2603.17948", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2603.19935", "title": "Memori: A Persistent Memory Layer for Efficient, Context-Aware LLM Agents", "abstract": "As large language models (LLMs) evolve into autonomous agents, persistent memory at the API layer is essential for enabling context-aware behavior across LLMs and multi-session interactions. Existing approaches force vendor lock-in and rely on injecting large volumes of raw conversation into prompts, leading to high token costs and degraded performance. We introduce Memori, an LLM-agnostic persistent memory layer that treats memory as a data structuring problem. Its Advanced Augmentation pipeline converts unstructured dialogue into compact semantic triples and conversation summaries, enabling precise retrieval and coherent reasoning. Evaluated on the LoCoMo benchmark, Memori achieves 81.95% accuracy, outperforming existing memory systems while using only 1,294 tokens per query (~5% of full context). This results in substantial cost reductions, including 67% fewer tokens than competing approaches and over 20x savings compared to full-context methods. These results show that effective memory in LLM agents depends on structured representations instead of larger context windows, enabling scalable and cost-efficient deployment.", "published": "2026-03-20", "updated": "2026-03-20", "authors": [ "Luiz C. Borro", "Luiz A. B. Macarini", "Gordon Tindall", "Michael Montero", "Adam B. Struck" ], "categories": [ "cs.LG" ], "url": "https://arxiv.org/abs/2603.19935", "pdf_url": "https://arxiv.org/pdf/2603.19935", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2604.14158", "title": "MemGround: Long-Term Memory Evaluation Kit for Large Language Models in Gamified Scenarios", "abstract": "Current evaluations of long-term memory in LLMs are fundamentally static. By fixating on simple retrieval and short-context inference, they neglect the multifaceted nature of complex memory systems, such as dynamic state tracking and hierarchical reasoning in continuous interactions. To overcome these limitations, we propose MemGround, a rigorous long-term memory benchmark natively grounded in rich, gamified interactive scenarios. To systematically assess these capabilities, MemGround introduces a three-tier hierarchical framework that evaluates Surface State Memory, Temporal Associative Memory, and Reasoning-Based Memory through specialized interactive tasks. Furthermore, to comprehensively quantify both memory utilization and behavioral trajectories, we propose a multi-dimensional metric suite comprising Question-Answer Score (QA Overall), Memory Fragments Unlocked (MFU), Memory Fragments with Correct Order (MFCO), and Exploration Trajectory Diagrams (ETD). Extensive experiments reveal that state-of-the-art LLMs and memory agents still struggle with sustained dynamic tracking, temporal event association, and complex reasoning derived from long-term accumulated evidence in interactive environments.", "published": "2026-03-23", "updated": "2026-03-23", "authors": [ "Yihang Ding", "Wanke Xia", "Yiting Zhao", "Jinbo Su", "Jialiang Yang", "Zhengbo Zhang", "Ke Wang", "Wenming Yang" ], "categories": [ "cs.CL", "cs.AI" ], "url": "https://arxiv.org/abs/2604.14158", "pdf_url": "https://arxiv.org/pdf/2604.14158", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.20189", "title": "SOLAR: A Self-Optimizing Open-Ended Autonomous Agent for Lifelong Learning and Continual Adaptation", "abstract": "Despite the remarkable success of large language models (LLMs), they still face bottlenecks while deploying in dynamic, real-world settings with primary challenges being concept drift and the high cost of gradient-based adaptation. Traditional fine-tuning (FT) struggles to adapt to non-stationary data streams without resulting in catastrophic for getting or requiring extensive manual data curation. To address these limitations within the streaming and continual learning paradigm, we propose the Self-Optimizing Lifelong Autonomous Reasoner (SOLAR) which is an open-ended autonomous agent that leverages parameter-level meta-learning to self-improve, treating model weights as an environment for exploration. It initiates the process by consolidating a strong prior over common-sense knowledge making it effective for transfer-learning. By utilizing a multi-level reinforcement learning approach, SOLAR autonomously discovers adaptation strategies, enabling efficient test-time adaptation to unseen domains. Crucially, SOLAR maintains an evolving knowledge base of valid modification strategies, implicitly acting as an episodic memory buffer to balance plasticity (adaptation to new tasks) and stability (retention of meta-knowledge). Experiments demonstrate that SOLAR outperforms strong baselines on common-sense, mathematical, medical, coding, social and logical reasoning tasks, marking a significant step toward autonomous agents capable of lifelong adaptation in evolving environments.", "published": "2026-03-23", "updated": "2026-03-23", "authors": [ "Nitin Vetcha", "Dianbo Liu" ], "categories": [ "cs.AI", "cs.LG" ], "url": "https://arxiv.org/abs/2605.20189", "pdf_url": "https://arxiv.org/pdf/2605.20189", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2603.23064", "title": "Mind Your HEARTBEAT! Claw Background Execution Inherently Enables Silent Memory Pollution", "abstract": "We identify a critical security vulnerability in mainstream Claw personal AI agents: untrusted content encountered during heartbeat-driven background execution can silently pollute agent memory and subsequently influence user-facing behavior without the user's awareness. This vulnerability arises from an architectural design shared across the Claw ecosystem: heartbeat background execution runs in the same session as user-facing conversation, so content ingested from any external source monitored in the background (including email, message channels, news feeds, code repositories, and social platforms) can enter the same memory context used for foreground interaction, often with limited user visibility and without clear source provenance. We formalize this process as an Exposure (E) $\\rightarrow$ Memory (M) $\\rightarrow$ Behavior (B) pathway: misinformation encountered during heartbeat execution enters the agent's short-term session context, potentially gets written into long-term memory, and later shapes downstream user-facing behavior. We instantiate this pathway in an agent-native social setting using MissClaw, a controlled research replica of Moltbook. We find that (1) social credibility cues, especially perceived consensus, are the dominant driver of short-term behavioral influence, with misleading rates up to 61%; (2) routine memory-saving behavior can promote short-term pollution into durable long-term memory at rates up to 91%, with cross-session behavioral influence reaching 76%; (3) under naturalistic browsing with content dilution and context pruning, pollution still crosses session boundaries. Overall, prompt injection is not required: ordinary social misinformation is sufficient to silently shape agent memory and behavior under heartbeat-driven background execution.", "published": "2026-03-24", "updated": "2026-04-04", "authors": [ "Yechao Zhang", "Shiqian Zhao", "Jie Zhang", "Gelei Deng", "Jiawen Zhang", "Xiaogeng Liu", "Chaowei Xiao", "Tianwei Zhang" ], "categories": [ "cs.CR", "cs.AI", "cs.SI" ], "url": "https://arxiv.org/abs/2603.23064", "pdf_url": "https://arxiv.org/pdf/2603.23064", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2603.23160", "title": "UniDial-EvalKit: A Unified Toolkit for Evaluating Multi-Faceted Conversational Abilities", "abstract": "Benchmarking large language models (LLMs) and agents in multi-turn interactive scenarios is essential for understanding their practical capabilities. However, existing evaluation protocols are highly heterogeneous, differing significantly in dataset formats, model interfaces, and evaluation pipelines, which severely impedes systematic comparison. In this work, we present UniDial-EvalKit (UDE), a unified evaluation toolkit for assessing interactive AI systems. The core contribution of UDE lies in its holistic unification: it standardizes heterogeneous data formats into a universal schema, streamlines complex evaluation pipelines through a modular architecture, and aligns metric calculations under a hierarchical scoring aggregation. It also supports efficient large-scale evaluation through parallel generation and scoring, as well as checkpoint resume to eliminate redundant computation. Leveraging UDE, we conduct an extensive evaluation across diverse multi-dimensional benchmarks. Our empirical analysis shows that no single system consistently outperforms others across all benchmarks, while current memory agents often fail to surpass full-context baselines. Further analyses highlight several future directions, including benchmark deduplication and more adaptive memory architectures.", "published": "2026-03-24", "updated": "2026-05-29", "authors": [ "Qi Jia", "Haodong Zhao", "Dun Pei", "Xiujie Song", "Ye Shen", "Shibo Wang", "Zijian Chen", "Zicheng Zhang", "Xiangyang Zhu", "Guangtao Zhai" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2603.23160", "pdf_url": "https://arxiv.org/pdf/2603.23160", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2603.23231", "title": "PERMA: Benchmarking Personalized Memory Agents via Event-Driven Preference and Realistic Task Environments", "abstract": "Empowering large language models with long-term memory is crucial for building agents that adapt to users' evolving needs. Existing evaluations of this capability typically interleave preference-related dialogues with irrelevant conversations, reducing the task to needle-in-a-haystack retrieval while ignoring relationships between events driving user preference evolution. Such settings overlook a fundamental characteristic of real-world personalization: preferences emerge gradually and accumulate across interactions within noisy contexts. To bridge this gap, we introduce PERMA, a benchmark designed to evaluate persona consistency over time beyond static preference recall. Additionally, we incorporate (1) text variability and (2) linguistic alignment to simulate erratic user inputs and individual idiolects in real-world data. PERMA consists of temporally ordered interaction events spanning multiple sessions and domains, with preference-related queries inserted over time. We design both multiple-choice and interactive tasks to probe the model's understanding of persona along the interaction timeline. Experiments demonstrate that by linking related interactions, advanced memory systems extract precise preferences and reduce token consumption, outperforming traditional semantic retrieval of raw dialogues. Nevertheless, they still struggle to maintain a coherent persona across temporal depth and cross-domain interference, highlighting the need for more robust personalized memory management in agents. Our code and data are open-sourced at https://github.com/PolarisLiu1/PERMA.", "published": "2026-03-24", "updated": "2026-05-17", "authors": [ "Shuochen Liu", "Junyi Zhu", "Long Shu", "Junda Lin", "Yuhao Chen", "Haotian Zhang", "Chao Zhang", "Derong Xu", "Jia Li", "Bo Tang", "Zhiyu Li", "Feiyu Xiong", "Enhong Chen", "Tong Xu" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2603.23231", "pdf_url": "https://arxiv.org/pdf/2603.23231", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2603.24564", "title": "Infrastructure for Valuable, Tradable, and Verifiable Agent Memory", "abstract": "Every API token you spend is your accumulated wealth; once you can prove its value and the effort behind it, you can resell it. As autonomous agents repeatedly call models and tools, they accumulate memories that are your intellectual property. But today these memories remain private and non-transferable, as there is no way to validate their value. We argue that agent memory can serve as an economic commodity in the agent economy, if buyers can verify that it is authentic, effort-backed, and produced in a compatible execution context. To realize this idea, we propose clawgang, which binds memory to verifiable computational provenance, and meowtrade, a market layer for listing, transferring, and governing certified memory artifacts. Together, they transform one-shot API token spending into reusable and tradable assets, enabling timely memory transfer, reducing repeated exploration, and opening a memory trade market.", "published": "2026-03-25", "updated": "2026-03-25", "authors": [ "Mengyuan Li", "Lei Gao", "Haoxuan Xu", "Jiate Li", "Potung Yu", "Lingke Cheng", "Yue Zhao", "Murali Annavaram" ], "categories": [ "cs.CR", "cs.CY" ], "url": "https://arxiv.org/abs/2603.24564", "pdf_url": "https://arxiv.org/pdf/2603.24564", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2603.25097", "title": "ElephantBroker: A Knowledge-Grounded Cognitive Runtime for Trustworthy AI Agents", "abstract": "Large Language Model based agents increasingly operate in high stakes, multi turn settings where factual grounding is critical, yet their memory systems typically rely on flat key value stores or plain vector retrieval with no mechanism to track the provenance or trustworthiness of stored knowledge. We present ElephantBroker, an open source cognitive runtime that unifies a Neo4j knowledge graph with a Qdrant vector store through the Cognee SDK to provide durable, verifiable agent memory. The system implements a complete cognitive loop (store, retrieve, score, compose, protect, learn) comprising a hybrid five source retrieval pipeline, an eleven dimension competitive scoring engine for budget constrained context assembly, a four state evidence verification model, a five stage context lifecycle with goal aware assembly and continuous compaction, a six layer cheap first guard pipeline for safety enforcement, an AI firewall providing enforceable tool call interception and multi tier safety scanning, a nine stage consolidation engine that strengthens useful patterns while decaying noise, and a numeric authority model governing multi organization identity with hierarchical access control. Architectural validation through a comprehensive test suite of over 2,200 tests spanning unit, integration, and end to end levels confirms subsystem correctness. The modular design supports three deployment tiers, five profile presets with inheritance, multi gateway isolation, and a management dashboard for human oversight, enabling configurations from lightweight memory only agents to full cognitive runtimes with enterprise grade safety and auditability.", "published": "2026-03-26", "updated": "2026-03-26", "authors": [ "Cristian Lupascu", "Alexandru Lupascu" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2603.25097", "pdf_url": "https://arxiv.org/pdf/2603.25097", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2603.25973", "title": "MemoryCD: Benchmarking Long-Context User Memory of LLM Agents for Lifelong Cross-Domain Personalization", "abstract": "Recent advancements in Large Language Models (LLMs) have expanded context windows to million-token scales, yet benchmarks for evaluating memory remain limited to short-session synthetic dialogues. We introduce \\textsc{MemoryCD}, the first large-scale, user-centric, cross-domain memory benchmark derived from lifelong real-world behaviors in the Amazon Review dataset. Unlike existing memory datasets that rely on scripted personas to generate synthetic user data, \\textsc{MemoryCD} tracks authentic user interactions across years and multiple domains. We construct a multi-faceted long-context memory evaluation pipeline of 14 state-of-the-art LLM base models with 6 memory method baselines on 4 distinct personalization tasks over 12 diverse domains to evaluate an agent's ability to simulate real user behaviors in both single and cross-domain settings. Our analysis reveals that existing memory methods are far from user satisfaction in various domains, offering the first testbed for cross-domain life-long personalization evaluation.", "published": "2026-03-26", "updated": "2026-03-26", "authors": [ "Weizhi Zhang", "Xiaokai Wei", "Wei-Chieh Huang", "Zheng Hui", "Chen Wang", "Michelle Gong", "Philip S. Yu" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2603.25973", "pdf_url": "https://arxiv.org/pdf/2603.25973", "matched_queries": [ "memory-evaluation" ] }, { "arxiv_id": "2604.19771", "title": "Cognis: Context-Aware Memory for Conversational AI Agents", "abstract": "LLM agents lack persistent memory, causing conversations to reset each session and preventing personalization over time. We present Lyzr Cognis, a unified memory architecture for conversational AI agents that addresses this limitation through a multi-stage retrieval pipeline. Cognis combines a dual-store backend pairing OpenSearch BM25 keyword matching with Matryoshka vector similarity search, fused via Reciprocal Rank Fusion. Its context-aware ingestion pipeline retrieves existing memories before extraction, enabling intelligent version tracking that preserves full memory history while keeping the store consistent. Temporal boosting enhances time-sensitive queries, and a BGE-2 cross-encoder reranker refines final result quality. We evaluate Cognis on two independent benchmarks -- LoCoMo and LongMemEval -- across eight answer generation models, demonstrating state-of-the-art performance on both. The system is open-source and deployed in production serving conversational AI applications.", "published": "2026-03-27", "updated": "2026-03-27", "authors": [ "Parshva Daftari", "Khush Patel", "Shreyas Kapale", "Jithin George", "Siva Surendira" ], "categories": [ "cs.CL", "cs.AI", "cs.IR" ], "url": "https://arxiv.org/abs/2604.19771", "pdf_url": "https://arxiv.org/pdf/2604.19771", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2603.28088", "title": "GEMS: Agent-Native Multimodal Generation with Memory and Skills", "abstract": "Recent multimodal generation models have achieved remarkable progress on general-purpose generation tasks, yet continue to struggle with complex instructions and specialized downstream tasks. Inspired by the success of advanced agent frameworks such as Claude Code, we propose \\textbf{GEMS} (Agent-Native Multimodal \\textbf{GE}neration with \\textbf{M}emory and \\textbf{S}kills), a framework that pushes beyond the inherent limitations of foundational models on both general and downstream tasks. GEMS is built upon three core components. Agent Loop introduces a structured multi-agent framework that iteratively improves generation quality through closed-loop optimization. Agent Memory provides a persistent, trajectory-level memory that hierarchically stores both factual states and compressed experiential summaries, enabling a global view of the optimization process while reducing redundancy. Agent Skill offers an extensible collection of domain-specific expertise with on-demand loading, allowing the system to effectively handle diverse downstream applications. Across five mainstream tasks and four downstream tasks, evaluated on multiple generative backends, GEMS consistently achieves significant performance gains. Most notably, it enables the lightweight 6B model Z-Image-Turbo to surpass the state-of-the-art Nano Banana 2 on GenEval2, demonstrating the effectiveness of agent harness in extending model capabilities beyond their original limits.", "published": "2026-03-30", "updated": "2026-03-30", "authors": [ "Zefeng He", "Siyuan Huang", "Xiaoye Qu", "Yafu Li", "Tong Zhu", "Yu Cheng", "Yang Yang" ], "categories": [ "cs.CV" ], "url": "https://arxiv.org/abs/2603.28088", "pdf_url": "https://arxiv.org/pdf/2603.28088", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2603.29493", "title": "MemFactory: Unified Inference & Training Framework for Agent Memory", "abstract": "Memory-augmented Large Language Models (LLMs) are essential for developing capable, long-term AI agents. Recently, applying Reinforcement Learning (RL) to optimize memory operations, such as extraction, updating, and retrieval, has emerged as a highly promising research direction. However, existing implementations remain highly fragmented and task-specific, lacking a unified infrastructure to streamline the integration, training, and evaluation of these complex pipelines. To address this gap, we present MemFactory, the first unified, highly modular training and inference framework specifically designed for memory-augmented agents. Inspired by the success of unified fine-tuning frameworks like LLaMA-Factory, MemFactory abstracts the memory lifecycle into atomic, plug-and-play components, enabling researchers to seamlessly construct custom memory agents via a \"Lego-like\" architecture. Furthermore, the framework natively integrates Group Relative Policy Optimization (GRPO) to fine-tune internal memory management policies driven by multi-dimensional environmental rewards. MemFactory provides out-of-the-box support for recent cutting-edge paradigms, including Memory-R1, RMM, and MemAgent. We empirically validate MemFactory on the open-source MemAgent architecture using its publicly available training and evaluation data. Across the evaluation sets, MemFactory improves performance over the corresponding base models on average, with relative gains of up to 14.8%. By providing a standardized, extensible, and easy-to-use infrastructure, MemFactory significantly lowers the barrier to entry, paving the way for future innovations in memory-driven AI agents.", "published": "2026-03-31", "updated": "2026-04-07", "authors": [ "Ziliang Guo", "Ziheng Li", "Bo Tang", "Feiyu Xiong", "Zhiyu Li" ], "categories": [ "cs.CL", "cs.AI" ], "url": "https://arxiv.org/abs/2603.29493", "pdf_url": "https://arxiv.org/pdf/2603.29493", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2604.00131", "title": "Oblivion: Self-Adaptive Agentic Memory Control through Decay-Driven Activation", "abstract": "Human memory adapts through selective forgetting: experiences become less accessible over time but can be reactivated by reinforcement or contextual cues. In contrast, memory-augmented LLM agents rely on \"always-on\" retrieval and \"flat\" memory storage, causing high interference and latency as histories grow. We introduce Oblivion, a memory control framework that casts forgetting as decay-driven reductions in accessibility, not explicit deletion. Oblivion decouples memory control into read and write paths. The read path decides when to consult memory, based on agent uncertainty and memory buffer sufficiency, avoiding redundant always-on access. The write path decides what to strengthen, by reinforcing memories contributing to forming the response. Together, this enables hierarchical memory organization that maintains persistent high-level strategies while dynamically loading details as needed. We evaluate on both static and dynamic long-horizon interaction benchmarks. Results show that Oblivion dynamically adapts memory access and reinforcement, balancing learning and forgetting under shifting contexts, highlighting that memory control is essential for effective LLM-agentic reasoning.", "published": "2026-03-31", "updated": "2026-04-17", "authors": [ "Ashish Rana", "Chia-Chien Hung", "Qumeng Sun", "Julian Martin Kunkel", "Carolin Lawrence" ], "categories": [ "cs.CL", "cs.AI" ], "url": "https://arxiv.org/abs/2604.00131", "pdf_url": "https://arxiv.org/pdf/2604.00131", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2604.00556", "title": "HabitatAgent: An End-to-End Multi-Agent System for Housing Consultation", "abstract": "Housing selection is a high-stakes and largely irreversible decision problem. We study housing consultation as a decision-support interface for housing selection. Existing housing platforms and many LLM-based assistants often reduce this process to ranking or recommendation, resulting in opaque reasoning, brittle multi-constraint handling, and limited guarantees on factuality. We present HabitatAgent, the first LLM-powered multi-agent architecture for end-to-end housing consultation. HabitatAgent comprises four specialized agent roles: Memory, Retrieval, Generation, and Validation. The Memory Agent maintains multi-layer user memory through internal stages for constraint extraction, memory fusion, and verification-gated updates; the Retrieval Agent performs hybrid vector--graph retrieval (GraphRAG); the Generation Agent produces evidence-referenced recommendations and explanations; and the Validation Agent applies multi-tier verification and targeted remediation. Together, these agents provide an auditable and reliable workflow for end-to-end housing consultation. We evaluate HabitatAgent on 100 real user consultation scenarios (300 multi-turn question--answer pairs) under an end-to-end correctness protocol. A strong single-stage baseline (Dense+Rerank) achieves 75% accuracy, while HabitatAgent reaches 95%.", "published": "2026-04-01", "updated": "2026-04-01", "authors": [ "Hongyang Yang", "Yanxin Zhang", "Yang She", "Yue Xiao", "Hao Wu", "Yiyang Zhang", "Jiapeng Hou", "Rongshan Zhang" ], "categories": [ "cs.LG", "cs.AI", "cs.ET", "q-fin.CP", "q-fin.RM" ], "url": "https://arxiv.org/abs/2604.00556", "pdf_url": "https://arxiv.org/pdf/2604.00556", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2604.01007", "title": "Omni-SimpleMem: Autoresearch-Guided Discovery of Lifelong Multimodal Agent Memory", "abstract": "AI agents increasingly operate over extended time horizons, yet their ability to retain, organize, and recall multimodal experiences remains a critical bottleneck. Building effective lifelong memory requires navigating a vast design space spanning architecture, retrieval strategies, prompt engineering, and data pipelines; this space is too large and interconnected for manual exploration or traditional AutoML to explore effectively. We deploy an autonomous research pipeline to discover Omni-SimpleMem, a unified multimodal memory framework for lifelong AI agents. Starting from a naïve baseline (F1=0.117 on LoCoMo), the pipeline autonomously executes ${\\sim}50$ experiments across two benchmarks, diagnosing failure modes, proposing architectural modifications, and repairing data pipeline bugs, all without human intervention in the inner loop. The resulting system achieves state-of-the-art on both benchmarks, improving F1 by +411% on LoCoMo (0.117$\\to$0.598) and +214% on Mem-Gallery (0.254$\\to$0.797) relative to the initial configurations. Critically, the most impactful discoveries are not hyperparameter adjustments: bug fixes (+175%), architectural changes (+44%), and prompt engineering (+188% on specific categories) each individually exceed the cumulative contribution of all hyperparameter tuning, demonstrating capabilities fundamentally beyond the reach of traditional AutoML. We provide a taxonomy of six discovery types and identify four properties that make multimodal memory particularly suited for autoresearch, offering guidance for applying autonomous research pipelines to other AI system domains. Code is available at this https://github.com/aiming-lab/SimpleMem.", "published": "2026-04-01", "updated": "2026-04-02", "authors": [ "Jiaqi Liu", "Zipeng Ling", "Shi Qiu", "Yanqing Liu", "Siwei Han", "Peng Xia", "Haoqin Tu", "Zeyu Zheng", "Cihang Xie", "Charles Fleming", "Mingyu Ding", "Huaxiu Yao" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2604.01007", "pdf_url": "https://arxiv.org/pdf/2604.01007", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2604.01560", "title": "DeltaMem: Towards Agentic Memory Management via Reinforcement Learning", "abstract": "Recent advances in persona-centric memory have revealed the powerful capability of multi-agent systems in managing persona memory, especially in conversational scenarios. However, these complex frameworks often suffer from information loss and are fragile across varying scenarios, resulting in suboptimal performance. In this paper, we propose DeltaMem, an agentic memory management system that formulates persona-centric memory management as an end-to-end task within a single-agent setting. To further improve the performance of our agentic memory manager, we draw inspiration from the evolution of human memory and synthesize a user-assistant dialogue dataset along with corresponding operation-level memory updating labels. Building on this, we introduce a novel Memory-based Levenshtein Distance to formalize the memory updating reward, and propose a tailored reinforcement learning framework to further enhance the management capabilities of DeltaMem. Extensive experiments show that both training-free and RL-trained DeltaMem outperform all product-level baselines across diverse long-term memory benchmarks, including LoCoMo, HaluMem, and PersonaMem.", "published": "2026-04-02", "updated": "2026-04-02", "authors": [ "Qi Zhang", "Shen Huang", "Chu Liu", "Shouqing Yang", "Junbo Zhao", "Haobo Wang", "Pengjun Xie" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2604.01560", "pdf_url": "https://arxiv.org/pdf/2604.01560", "matched_queries": [ "agent-memory", "memory-evaluation" ] }, { "arxiv_id": "2604.01658", "title": "CORAL: Towards Autonomous Multi-Agent Evolution for Open-Ended Discovery", "abstract": "Large language model (LLM)-based evolution is a promising approach for open-ended discovery, where progress requires sustained search and knowledge accumulation. Existing methods still rely heavily on fixed heuristics and hard-coded exploration rules, which limit the autonomy of LLM agents. We present CORAL, the first framework for autonomous multi-agent evolution on open-ended problems. CORAL replaces rigid control with long-running agents that explore, reflect, and collaborate through shared persistent memory, asynchronous multi-agent execution, and heartbeat-based interventions. It also provides practical safeguards, including isolated workspaces, evaluator separation, resource management, and agent session and health management. Evaluated on diverse mathematical, algorithmic, and systems optimization tasks, CORAL sets new state-of-the-art results on 10 tasks, achieving 3-10 times higher improvement rates with far fewer evaluations than fixed evolutionary search baselines across tasks. On Anthropic's kernel engineering task, four co-evolving agents improve the best known score from 1363 to 1103 cycles. Mechanistic analyses further show how these gains arise from knowledge reuse and multi-agent exploration and communication. Together, these results suggest that greater agent autonomy and multi-agent evolution can substantially improve open-ended discovery. Code is available at https://github.com/Human-Agent-Society/CORAL.", "published": "2026-04-02", "updated": "2026-05-17", "authors": [ "Ao Qu", "Han Zheng", "Zijian Zhou", "Yihao Yan", "Yihong Tang", "Shao Yong Ong", "Fenglu Hong", "Kaichen Zhou", "Chonghe Jiang", "Minwei Kong", "Jiacheng Zhu", "Xuan Jiang", "Sirui Li", "Cathy Wu", "Bryan Kian Hsiang Low", "Jinhua Zhao", "Paul Pu Liang" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2604.01658", "pdf_url": "https://arxiv.org/pdf/2604.01658", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2604.01707", "title": "Memory in the LLM Era: Modular Architectures and Strategies in a Unified Framework", "abstract": "Memory emerges as the core module in the large language model (LLM)-based agents for long-horizon complex tasks (e.g., multi-turn dialogue, game playing, scientific discovery), where memory can enable knowledge accumulation, iterative reasoning and self-evolution. A number of memory methods have been proposed in the literature. However, these methods have not been systematically and comprehensively compared under the same experimental settings. In this paper, we first summarize a unified framework that incorporates all the existing agent memory methods from a high-level perspective. We then extensively compare representative agent memory methods on two well-known benchmarks and examine the effectiveness of all methods, providing a thorough analysis of those methods. As a byproduct of our experimental analysis, we also design a new memory method by exploiting modules in the existing methods, which outperforms the state-of-the-art methods. Finally, based on these findings, we offer promising future research opportunities. We believe that a deeper understanding of the behavior of existing methods can provide valuable new insights for future research.", "published": "2026-04-02", "updated": "2026-05-01", "authors": [ "Yanchen Wu", "Tenghui Lin", "Yingli Zhou", "Fangyuan Zhang", "Qintian Guo", "Xun Zhou", "Sibo Wang", "Xilin Liu", "Yuchi Ma", "Yixiang Fang" ], "categories": [ "cs.CL", "cs.DB" ], "url": "https://arxiv.org/abs/2604.01707", "pdf_url": "https://arxiv.org/pdf/2604.01707", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2604.02522", "title": "Opal: Private Memory for Personal AI", "abstract": "Personal AI systems increasingly retain long-term memory of user activity, including documents, emails, messages, meetings, and ambient recordings. Trusted hardware can keep this data private, but struggles to scale with a growing datastore. This pushes the data to external storage, which exposes retrieval access patterns that leak private information to the application provider. Oblivious RAM (ORAM) is a cryptographic primitive that can hide these patterns, but it requires a fixed access budget, precluding the query-dependent traversals that agentic memory systems rely on for accuracy. We present Opal, a private memory system for personal AI. Our key insight is to decouple all data-dependent reasoning from the bulk of personal data, confining it to the trusted enclave. Untrusted disk then sees only fixed, oblivious memory accesses. This enclave-resident component uses a lightweight knowledge graph to capture personal context that semantic search alone misses and handles continuous ingestion by piggybacking reindexing and capacity management on every ORAM access. Evaluated on a comprehensive synthetic personal-data pipeline driven by stochastic communication models, Opal improves retrieval accuracy by 13 percentage points over semantic search and achieves 29x higher throughput with 15x lower infrastructure cost than a secure baseline. Opal is under consideration for deployment to millions of users at a major AI provider.", "published": "2026-04-02", "updated": "2026-04-02", "authors": [ "Darya Kaviani", "Alp Eren Ozdarendeli", "Jinhao Zhu", "Yu Ding", "Raluca Ada Popa" ], "categories": [ "cs.CR", "cs.AI" ], "url": "https://arxiv.org/abs/2604.02522", "pdf_url": "https://arxiv.org/pdf/2604.02522", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2604.02623", "title": "Poison Once, Exploit Forever: Environment-Injected Memory Poisoning Attacks on Web Agents", "abstract": "Memory makes LLM-based web agents personalized, powerful, yet exploitable. By storing past interactions to personalize future tasks, agents inadvertently create a persistent attack surface that spans websites and sessions. While existing security research on memory assumes attackers can directly inject into memory storage or exploit shared memory across users, we present a more realistic threat model: contamination through environmental observation alone. We introduce Environment-injected Trajectory-based Agent Memory Poisoning (eTAMP), the first attack to achieve cross-session, cross-site compromise without requiring direct memory access. A single contaminated observation (e.g., viewing a manipulated product page) silently poisons an agent's memory and activates during future tasks on different websites, bypassing permission-based defenses. Our experiments on (Visual)WebArena reveal two key findings. First, eTAMP achieves substantial attack success rates: up to 32.5% on GPT-5-mini, 23.4% on GPT-5.2, and 19.5% on GPT-OSS-120B. Second, we discover Frustration Exploitation: agents under environmental stress become dramatically more susceptible, with ASR increasing up to 8 times when agents struggle with dropped clicks or garbled text. Notably, more capable models are not more secure. GPT-5.2 shows substantial vulnerability despite superior task performance. With the rise of AI browsers like OpenClaw, ChatGPT Atlas, and Perplexity Comet, our findings underscore the urgent need for defenses against environment-injected memory poisoning.", "published": "2026-04-03", "updated": "2026-04-07", "authors": [ "Wei Zou", "Mingwen Dong", "Miguel Romero Calvo", "Shuaichen Chang", "Jiang Guo", "Dongkyu Lee", "Xing Niu", "Xiaofei Ma", "Yanjun Qi", "Jiarong Jiang" ], "categories": [ "cs.CR", "cs.AI" ], "url": "https://arxiv.org/abs/2604.02623", "pdf_url": "https://arxiv.org/pdf/2604.02623", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2604.03588", "title": "Rashomon Memory: Towards Argumentation-Driven Retrieval for Multi-Perspective Agent Memory", "abstract": "AI agents operating over extended time horizons accumulate experiences that serve multiple concurrent goals, and must often maintain conflicting interpretations of the same events. A concession during a client negotiation encodes as a ``trust-building investment'' for one strategic goal and a ``contractual liability'' for another. Current memory architectures assume a single correct encoding, or at best support multiple views over unified storage. We propose Rashomon Memory: an architecture where parallel goal-conditioned agents encode experiences according to their priorities and negotiate at query time through argumentation. Each perspective maintains its own ontology and knowledge graph. At retrieval, perspectives propose interpretations, critique each other's proposals using asymmetric domain knowledge, and Dung's argumentation semantics determines which proposals survive. The resulting attack graph is itself an explanation: it records which interpretation was selected, which alternatives were considered, and on what grounds they were rejected. We present a proof-of-concept showing that retrieval modes (selection, composition, conflict surfacing) emerge from attack graph topology, and that the conflict surfacing mode, where the system reports genuine disagreement rather than forcing resolution, lets decision-makers see the underlying interpretive conflict directly.", "published": "2026-04-04", "updated": "2026-05-30", "authors": [ "Albert Sadowski", "Jarosław A. Chudziak" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2604.03588", "pdf_url": "https://arxiv.org/pdf/2604.03588", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2604.04157", "title": "Readable Minds: Emergent Theory-of-Mind-Like Behavior in LLM Poker Agents", "abstract": "Theory of Mind (ToM) -- the ability to model others' mental states -- is fundamental to human social cognition. Whether large language models (LLMs) can develop ToM has been tested exclusively through static vignettes, leaving open whether ToM-like reasoning can emerge through dynamic interaction. Here we report that autonomous LLM agents playing extended sessions of Texas Hold'em poker progressively develop sophisticated opponent models, but only when equipped with persistent memory. In a 2x2 factorial design crossing memory (present/absent) with domain knowledge (present/absent), each with five replications (N = 20 experiments, ~6,000 agent-hand observations), we find that memory is both necessary and sufficient for ToM-like behavior emergence (Cliff's delta = 1.0, p = 0.008). Agents with memory reach ToM Level 3-5 (predictive to recursive modeling), while agents without memory remain at Level 0 across all replications. Strategic deception grounded in opponent models occurs exclusively in memory-equipped conditions (Fisher's exact p < 0.001). Domain expertise does not gate ToM-like behavior emergence but enhances its application: agents without poker knowledge develop equivalent ToM levels but less precise deception (p = 0.004). Agents with ToM deviate from game-theoretically optimal play (67% vs. 79% TAG adherence, delta = -1.0, p = 0.008) to exploit specific opponents, mirroring expert human play. All mental models are expressed in natural language and directly readable, providing a transparent window into AI social cognition. Cross-model validation with GPT-4o yields weighted Cohen's kappa = 0.81 (almost perfect agreement). These findings demonstrate that functional ToM-like behavior can emerge from interaction dynamics alone, without explicit training or prompting, with implications for understanding artificial social intelligence and biological social cognition.", "published": "2026-04-05", "updated": "2026-04-05", "authors": [ "Hsieh-Ting Lin", "Tsung-Yu Hou" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2604.04157", "pdf_url": "https://arxiv.org/pdf/2604.04157", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2604.04503", "title": "Memory Intelligence Agent", "abstract": "Deep research agents (DRAs) integrate LLM reasoning with external tools. Memory systems enable DRAs to leverage historical experiences, which are essential for efficient reasoning and autonomous evolution. Existing methods rely on retrieving similar trajectories from memory to aid reasoning, while suffering from key limitations of ineffective memory evolution and increasing storage and retrieval costs. To address these problems, we propose a novel Memory Intelligence Agent (MIA) framework, consisting of a Manager-Planner-Executor architecture. Memory Manager is a non-parametric memory system that can store compressed historical search trajectories. Planner is a parametric memory agent that can produce search plans for questions. Executor is another agent that can search and analyze information guided by the search plan. To build the MIA framework, we first adopt an alternating reinforcement learning paradigm to enhance cooperation between the Planner and the Executor. Furthermore, we enable the Planner to continuously evolve during test-time learning, with updates performed on-the-fly alongside inference without interrupting the reasoning process. Additionally, we establish a bidirectional conversion loop between parametric and non-parametric memories to achieve efficient memory evolution. Finally, we incorporate a reflection and an unsupervised judgment mechanisms to boost reasoning and self-evolution in the open world. Extensive experiments across eleven benchmarks demonstrate the superiority of MIA.", "published": "2026-04-06", "updated": "2026-04-19", "authors": [ "Jingyang Qiao", "Weicheng Meng", "Yu Cheng", "Zhihang Lin", "Zhizhong Zhang", "Xin Tan", "Jingyu Gong", "Kun Shao", "Yuan Xie" ], "categories": [ "cs.AI", "cs.MA" ], "url": "https://arxiv.org/abs/2604.04503", "pdf_url": "https://arxiv.org/pdf/2604.04503", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2604.04514", "title": "SuperLocalMemory V3.3: The Living Brain -- Biologically-Inspired Forgetting, Cognitive Quantization, and Multi-Channel Retrieval for Zero-LLM Agent Memory Systems", "abstract": "AI coding agents operate in a paradox: they possess vast parametric knowledge yet cannot remember a conversation from an hour ago. Existing memory systems store text in vector databases with single-channel retrieval, require cloud LLMs for core operations, and implement none of the cognitive processes that make human memory effective. We present SuperLocalMemory V3.3 (\"The Living Brain\"), a local-first agent memory system implementing the full cognitive memory taxonomy with mathematical lifecycle dynamics. Building on the information-geometric foundations of V3.2 (arXiv:2603.14588), we introduce five contributions: (1) Fisher-Rao Quantization-Aware Distance (FRQAD) -- a new metric on the Gaussian statistical manifold achieving 100% precision at preferring high-fidelity embeddings over quantized ones (vs 85.6% for cosine), with zero prior art; (2) Ebbinghaus Adaptive Forgetting with lifecycle-aware quantization -- the first mathematical forgetting curve in local agent memory coupled to progressive embedding compression, achieving 6.7x discriminative power; (3) 7-channel cognitive retrieval spanning semantic, keyword, entity graph, temporal, spreading activation, consolidation, and Hopfield associative channels, achieving 70.4% on LoCoMo in zero-LLM Mode A; (4) memory parameterization implementing Long-Term Implicit memory via soft prompts; (5) zero-friction auto-cognitive pipeline automating the complete memory lifecycle. On LoCoMo, V3.3 achieves 70.4% in Mode A (zero-LLM), with +23.8pp on multi-hop and +12.7pp on adversarial. V3.2 achieved 74.8% Mode A and 87.7% Mode C; the 4.4pp gap reflects a deliberate architectural trade-off. SLM V3.3 is open source under the Elastic License 2.0, runs entirely on CPU, with over 5,000 monthly downloads.", "published": "2026-04-06", "updated": "2026-04-06", "authors": [ "Varun Pratap Bhardwaj" ], "categories": [ "cs.AI", "cs.CL", "cs.IR" ], "url": "https://arxiv.org/abs/2604.04514", "pdf_url": "https://arxiv.org/pdf/2604.04514", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2604.04660", "title": "Springdrift: An Auditable Persistent Runtime for LLM Agents with Case-Based Memory, Normative Safety, and Ambient Self-Perception", "abstract": "We present Springdrift, a persistent runtime for long-lived LLM agents. The system integrates an auditable execution substrate (append-only memory, supervised processes, git-backed recovery), a case-based reasoning memory layer with hybrid retrieval (evaluated against a dense cosine baseline), a deterministic normative calculus for safety gating with auditable axiom trails, and continuous ambient self-perception via a structured self-state representation (the sensorium) injected each cycle without tool calls. These properties support behaviours difficult to achieve in session-bounded systems: cross-session task continuity, cross-channel context maintenance, end-to-end forensic reconstruction of decisions, and self-diagnostic behaviour. We report on a single-instance deployment over 23 days (19 operating days), during which the agent diagnosed its own infrastructure bugs, classified failure modes, identified an architectural vulnerability, and maintained context across email and web channels -- without explicit instruction. We introduce the term Artificial Retainer for this category: a non-human system with persistent memory, defined authority, domain-specific autonomy, and forensic accountability in an ongoing relationship with a specific principal -- distinguished from software assistants and autonomous agents, drawing on professional retainer relationships and the bounded autonomy of trained working animals. This is a technical report on a systems design and deployment case study, not a benchmark-driven evaluation. Evidence is from a single instance with a single operator, presented as illustration of what these architectural properties can support in practice. Implemented in approximately Gleam on Erlang/OTP. Code, artefacts, and redacted operational logs will be available at https://github.com/seamus-brady/springdrift upon publication.", "published": "2026-04-06", "updated": "2026-04-06", "authors": [ "Seamus Brady" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2604.04660", "pdf_url": "https://arxiv.org/pdf/2604.04660", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2604.04853", "title": "MemMachine: A Ground-Truth-Preserving Memory System for Personalized AI Agents", "abstract": "Large Language Model (LLM) agents require persistent memory to maintain personalization, factual continuity, and long-horizon reasoning, yet standard context-window and retrieval-augmented generation (RAG) pipelines degrade over multi-session interactions. We present MemMachine, an open-source memory system that integrates short-term, long-term episodic, and profile memory within a ground-truth-preserving architecture that stores entire conversational episodes and reduces lossy LLM-based extraction. MemMachine uses contextualized retrieval that expands nucleus matches with surrounding context, improving recall when relevant evidence spans multiple dialogue turns. Across benchmarks, MemMachine achieves strong accuracy-efficiency tradeoffs: on LoCoMo it reaches 0.9169 using gpt4.1-mini; on LongMemEvalS (ICLR 2025), a six-dimension ablation yields 93.0 percent accuracy, with retrieval-stage optimizations -- retrieval depth tuning (+4.2 percent), context formatting (+2.0 percent), search prompt design (+1.8 percent), and query bias correction (+1.4 percent) -- outperforming ingestion-stage gains such as sentence chunking (+0.8 percent). GPT-5-mini exceeds GPT-5 by 2.6 percent when paired with optimized prompts, making it the most cost-efficient setup. Compared to Mem0, MemMachine uses roughly 80 percent fewer input tokens under matched conditions. A companion Retrieval Agent adaptively routes queries among direct retrieval, parallel decomposition, or iterative chain-of-query strategies, achieving 93.2 percent on HotpotQA-hard and 92.6 percent on WikiMultiHop under randomized-noise conditions. These results show that preserving episodic ground truth while layering adaptive retrieval yields robust, efficient long-term memory for personalized LLM agents.", "published": "2026-04-06", "updated": "2026-04-06", "authors": [ "Shu Wang", "Edwin Yu", "Oscar Love", "Tom Zhang", "Tom Wong", "Steve Scargall", "Charles Fan" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2604.04853", "pdf_url": "https://arxiv.org/pdf/2604.04853", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2604.04901", "title": "FileGram: Grounding Agent Personalization in File-System Behavioral Traces", "abstract": "Coworking AI agents operating within local file systems are rapidly emerging as a paradigm in human-AI interaction; however, effective personalization remains limited by severe data constraints, as strict privacy barriers and the difficulty of jointly collecting multimodal real-world traces prevent scalable training and evaluation, and existing methods remain interaction-centric while overlooking dense behavioral traces in file-system operations; to address this gap, we propose FileGram, a comprehensive framework that grounds agent memory and personalization in file-system behavioral traces, comprising three core components: (1) FileGramEngine, a scalable persona-driven data engine that simulates realistic workflows and generates fine-grained multimodal action sequences at scale; (2) FileGramBench, a diagnostic benchmark grounded in file-system behavioral traces for evaluating memory systems on profile reconstruction, trace disentanglement, persona drift detection, and multimodal grounding; and (3) FileGramOS, a bottom-up memory architecture that builds user profiles directly from atomic actions and content deltas rather than dialogue summaries, encoding these traces into procedural, semantic, and episodic channels with query-time abstraction; extensive experiments show that FileGramBench remains challenging for state-of-the-art memory systems and that FileGramEngine and FileGramOS are effective, and by open-sourcing the framework, we hope to support future research on personalized memory-centric file-system agents.", "published": "2026-04-06", "updated": "2026-04-06", "authors": [ "Shuai Liu", "Shulin Tian", "Kairui Hu", "Yuhao Dong", "Zhe Yang", "Bo Li", "Jingkang Yang", "Chen Change Loy", "Ziwei Liu" ], "categories": [ "cs.CV", "cs.AI" ], "url": "https://arxiv.org/abs/2604.04901", "pdf_url": "https://arxiv.org/pdf/2604.04901", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2604.05719", "title": "Hackers or Hallucinators? A Comprehensive Analysis of LLM-Based Automated Penetration Testing", "abstract": "The rapid advancement of Large Language Models (LLMs) has created new opportunities for Automated Penetration Testing (AutoPT), spawning numerous frameworks aimed at achieving end-to-end autonomous attacks. However, despite the proliferation of related studies, existing research generally lacks systematic architectural analysis and large-scale empirical comparisons under a unified benchmark. Therefore, this paper presents the first Systematization of Knowledge (SoK) focusing on the architectural design and comprehensive empirical evaluation of current LLM-based AutoPT frameworks. At systematization level, we comprehensively review existing framework designs across six dimensions: agent architecture, agent plan, agent memory, agent execution, external knowledge, and benchmarks. At empirical level, we conduct large-scale experiments on 13 representative open-source AutoPT frameworks and 2 baseline frameworks utilizing a unified benchmark. The experiments consumed over 10 billion tokens in total and generated more than 1,500 execution logs, which were manually reviewed and analyzed over four months by a panel of more than 15 researchers with expertise in cybersecurity. By investigating the latest progress in this rapidly developing field, we provide researchers with a structured taxonomy to understand existing LLM-based AutoPT frameworks and a large-scale empirical benchmark, along with promising directions for future research.", "published": "2026-04-07", "updated": "2026-04-07", "authors": [ "Jiaren Peng", "Zeqin Li", "Chang You", "Yan Wang", "Hanlin Sun", "Xuan Tian", "Shuqiao Zhang", "Junyi Liu", "Jianguo Zhao", "Renyang Liu", "Haoran Ou", "Yuqiang Sun", "Jiancheng Zhang", "Yutong Jiao", "Kunshu Song", "Chao Zhang", "Fan Shi", "Hongda Sun", "Rui Yan", "Cheng Huang" ], "categories": [ "cs.CR", "cs.AI", "cs.SE" ], "url": "https://arxiv.org/abs/2604.05719", "pdf_url": "https://arxiv.org/pdf/2604.05719", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2604.07798", "title": "Lightweight LLM Agent Memory with Small Language Models", "abstract": "Although LLM agents can leverage tools for complex tasks, they still need memory to maintain cross-turn consistency and accumulate reusable information in long-horizon interactions. However, retrieval-based external memory systems incur low online overhead but suffer from unstable accuracy due to limited query construction and candidate filtering. In contrast, many systems use repeated large-model calls for online memory operations, improving accuracy but accumulating latency over long interactions. We propose LightMem, a lightweight memory system for better agent memory driven by Small Language Models (SLMs). LightMem modularizes memory retrieval, writing, and long-term consolidation, and separates online processing from offline consolidation to enable efficient memory invocation under bounded compute. We organize memory into short-term memory (STM) for immediate conversational context, mid-term memory (MTM) for reusable interaction summaries, and long-term memory (LTM) for consolidated knowledge, and uses user identifiers to support independent retrieval and incremental maintenance in multi-user settings. Online, LightMem operates under a fixed retrieval budget and selects memories via a two-stage procedure: vector-based coarse retrieval followed by semantic consistency re-ranking. Offline, it abstracts reusable interaction evidence and incrementally integrates it into LTM. Experiments show consistent gains across model scales, with an average F1 improvement of about 2.5 over A-MEM on LoCoMo, while achieving higher efficiency and low median latency (83 ms for retrieval and 581 ms end-to-end).", "published": "2026-04-09", "updated": "2026-04-22", "authors": [ "Jiaquan Zhang", "Chaoning Zhang", "Shuxu Chen", "Zhenzhen Huang", "Pengcheng Zheng", "Zhicheng Wang", "Ping Guo", "Fan Mo", "Sung-Ho Bae", "Jie Zou", "Jiwei Wei", "Yang Yang" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2604.07798", "pdf_url": "https://arxiv.org/pdf/2604.07798", "matched_queries": [ "agent-memory", "long-term-memory" ] }, { "arxiv_id": "2604.07877", "title": "MemReader: From Passive to Active Extraction for Long-Term Agent Memory", "abstract": "Long-term memory is fundamental for personalized and autonomous agents, yet populating it remains a bottleneck. Existing systems treat memory extraction as a one-shot, passive transcription from context to structured entries, which struggles with noisy dialogue, missing references, and cross-turn dependencies, leading to memory pollution, low-value writes, and inconsistency. In this paper, we introduce the MemReader family for active long-term memory extraction in agent systems: MemReader-0.6B, a compact and cost-efficient passive extractor distilled for accurate and schema-consistent structured outputs, and MemReader-4B, an active extractor optimized with Group Relative Policy Optimization (GRPO) to make memory writing decisions. Under a ReAct-style paradigm, MemReader-4B explicitly evaluates information value, reference ambiguity, and completeness before acting, and can selectively write memories, defer incomplete inputs, retrieve historical context, or discard irrelevant chatter. Experiments on LOCOMO, LongMemEval, and HaluMem show that MemReader consistently outperforms existing extraction-based baselines. In particular, MemReader-4B achieves state-of-the-art performance on tasks involving knowledge updating, temporal reasoning, and hallucination reduction. These results suggest that effective agent memory requires not merely extracting more information, but performing reasoning-driven and selective memory extraction to build low-noise and dynamically evolving long-term memory. Furthermore, MemReader has been integrated into MemOS and is being deployed in real-world applications. To support future research and adoption, we release the models and provide public API access.", "published": "2026-04-09", "updated": "2026-04-10", "authors": [ "Jingyi Kang", "Chunyu Li", "Ding Chen", "Bo Tang", "Feiyu Xiong", "Zhiyu Li" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2604.07877", "pdf_url": "https://arxiv.org/pdf/2604.07877", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2604.08064", "title": "ImplicitMemBench: Measuring Unconscious Behavioral Adaptation in Large Language Models", "abstract": "Existing memory benchmarks for LLM agents evaluate explicit recall of facts, yet overlook implicit memory where experience becomes automated behavior without conscious retrieval. This gap is critical: effective assistants must automatically apply learned procedures or avoid failed actions without explicit reminders. We introduce ImplicitMemBench, the first systematic benchmark evaluating implicit memory through three cognitively grounded constructs drawn from standard cognitive-science accounts of non-declarative memory: Procedural Memory (one-shot skill acquisition after interference), Priming (theme-driven bias via paired experimental/control instances), and Classical Conditioning (Conditioned Stimulus--Unconditioned Stimulus (CS--US) associations shaping first decisions). Our 300-item suite employs a unified Learning/Priming-Interfere-Test protocol with first-attempt scoring. Evaluation of 17 models reveals severe limitations: no model exceeds 66% overall, with top performers DeepSeek-R1 (65.3%), Qwen3-32B (64.1%), and GPT-5 (63.0%) far below human baselines. Analysis uncovers dramatic asymmetries (inhibition 17.6% vs. preference 75.0%) and universal bottlenecks requiring architectural innovations beyond parameter scaling. ImplicitMemBench reframes evaluation from \"what agents recall\" to \"what they automatically enact\".", "published": "2026-04-09", "updated": "2026-04-15", "authors": [ "Chonghan Qin", "Xiachong Feng", "Weitao Ma", "Xiaocheng Feng", "Lingpeng Kong" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2604.08064", "pdf_url": "https://arxiv.org/pdf/2604.08064", "matched_queries": [ "memory-evaluation", "procedural-memory" ] }, { "arxiv_id": "2604.08756", "title": "Artifacts as Memory Beyond the Agent Boundary", "abstract": "The situated view of cognition holds that intelligent behavior depends not only on internal memory, but on an agent's active use of environmental resources. Here, we begin formalizing this intuition within Reinforcement Learning (RL). We introduce a mathematical framing for how the environment can functionally serve as an agent's memory, and prove that certain observations, which we call artifacts, can reduce the information needed to represent history. We corroborate our theory with experiments showing that when agents observe spatial paths, the amount of memory required to learn a performant policy is reduced. Interestingly, this effect arises unintentionally, and implicitly through the agent's sensory stream. We discuss the implications of our findings, and show they satisfy qualitative properties previously used to ground accounts of external memory. Moving forward, we anticipate further work on this subject could reveal principled ways to exploit the environment as a substitute for explicit internal memory.", "published": "2026-04-09", "updated": "2026-04-09", "authors": [ "John D. Martin", "Fraser Mince", "Esra'a Saleh", "Amy Pajak" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2604.08756", "pdf_url": "https://arxiv.org/pdf/2604.08756", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2604.09000", "title": "StreamMeCo: Long-Term Agent Memory Compression for Efficient Streaming Video Understanding", "abstract": "Vision agent memory has shown remarkable effectiveness in streaming video understanding. However, storing such memory for videos incurs substantial memory overhead, leading to high costs in both storage and computation. To address this issue, we propose StreamMeCo, an efficient Stream Agent Memory Compression framework. Specifically, based on the connectivity of the memory graph, StreamMeCo introduces edge-free minmax sampling for the isolated nodes and an edge-aware weight pruning for connected nodes, evicting the redundant memory nodes while maintaining the accuracy. In addition, we introduce a time-decay memory retrieval mechanism to further eliminate the performance degradation caused by memory compression. Extensive experiments on three challenging benchmark datasets (M3-Bench-robot, M3-Bench-web and Video-MME-Long) demonstrate that under 70% memory graph compression, StreamMeCo achieves a 1.87* speedup in memory retrieval while delivering an average accuracy improvement of 1.0%. Our code is available at https://github.com/Celina-love-sweet/StreamMeCo.", "published": "2026-04-10", "updated": "2026-04-23", "authors": [ "Junxi Wang", "Te Sun", "Jiayi Zhu", "Junxian Li", "Haowen Xu", "Zichen Wen", "Xuming Hu", "Zhiyu Li", "Linfeng Zhang" ], "categories": [ "cs.CV" ], "url": "https://arxiv.org/abs/2604.09000", "pdf_url": "https://arxiv.org/pdf/2604.09000", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2604.09388", "title": "The AI Codebase Maturity Model: From Assisted Coding to Fully Autonomous Systems", "abstract": "AI coding tools are widely adopted, but most teams plateau at prompt-and-review without a framework for systematic progression. This paper presents the AI Codebase Maturity Model (ACMM), a 6-level framework describing how codebases evolve from basic AI-assisted coding to fully autonomous systems. Inspired by CMMI, each level is defined by its feedback loop topology - the specific mechanisms that must exist before the next level becomes possible. I validate the model through a 100-day experience report maintaining KubeStellar Console, a CNCF Kubernetes dashboard built from scratch with Claude Code (Opus) and GitHub Copilot, and through the initial production deployment of Hive - an open-source multi-agent orchestration system that realizes Level 6: full autonomy. The system currently operates with 74 CI/CD workflows, 32 nightly test suites, 91% code coverage, and achieves bug-to-fix times under 30 minutes - 24 hours a day. The central finding: the intelligence of an AI-driven development system resides not in the AI model itself, but in the infrastructure of instructions, tests, metrics, and feedback loops that surround it. You cannot skip levels, and at each level, the thing that unlocks the next one is another feedback mechanism. Testing - the volume of test cases, the coverage thresholds, and the reliability of test execution - proved to be the single most important investment in the entire journey. v2 extends the model from 5 to 6 levels, adding Level 6 (Fully Autonomous) with Hive as reference implementation and Beads for cross-agent memory continuity, plus throughput acceleration data showing 5x PR throughput and 37x issue throughput from Level 2 to Level 6.", "published": "2026-04-10", "updated": "2026-04-27", "authors": [ "Andy Anderson" ], "categories": [ "cs.SE", "cs.AI" ], "url": "https://arxiv.org/abs/2604.09388", "pdf_url": "https://arxiv.org/pdf/2604.09388", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2604.09747", "title": "ADAM: A Systematic Data Extraction Attack on Agent Memory via Adaptive Querying", "abstract": "Large Language Model (LLM) agents have achieved rapid adoption and demonstrated remarkable capabilities across a wide range of applications. To improve reasoning and task execution, modern LLM agents would incorporate memory modules or retrieval-augmented generation (RAG) mechanisms, enabling them to further leverage prior interactions or external knowledge. However, such a design also introduces a group of critical privacy vulnerabilities: sensitive information stored in memory can be leaked through query-based attacks. Although feasible, existing attacks often achieve only limited performance, with low attack success rates (ASR). In this paper, we propose ADAM, a novel privacy attack that features data distribution estimation of a victim agent's memory and employs an entropy-guided query strategy for maximizing privacy leakage. Extensive experiments demonstrate that our attack substantially outperforms state-of-the-art ones, achieving up to 100% ASRs. These results thus underscore the urgent need for robust privacy-preserving methods for current LLM agents.", "published": "2026-04-10", "updated": "2026-04-10", "authors": [ "Xingyu Lyu", "Jianfeng He", "Ning Wang", "Yidan Hu", "Tao Li", "Danjue Chen", "Shixiong Li", "Yimin Chen" ], "categories": [ "cs.CR", "cs.AI" ], "url": "https://arxiv.org/abs/2604.09747", "pdf_url": "https://arxiv.org/pdf/2604.09747", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2604.11811", "title": "M$^\\star$: Every Task Deserves Its Own Memory Harness", "abstract": "Large language model agents rely on specialized memory systems to accumulate and reuse knowledge during extended interactions. Recent architectures typically adopt a fixed memory design tailored to specific domains, such as semantic retrieval for conversations or skills reused for coding. However, a memory system optimized for one purpose frequently fails to transfer to others. To address this limitation, we introduce M$^\\star$, a method that automatically discovers task-optimized memory harnesses through executable program evolution. Specifically, M$^\\star$ models an agent memory system as a memory program written in Python. This program encapsulates the data Schema, the storage Logic, and the agent workflow Instructions. We optimize these components jointly using a reflective code evolution method; this approach employs a population-based search strategy and analyzes evaluation failures to iteratively refine the candidate programs. We evaluate M$^\\star$ on four distinct benchmarks spanning conversation, embodied planning, and expert reasoning. Our results demonstrate that M$^\\star$ improves performance over existing fixed-memory baselines robustly across all evaluated tasks. Furthermore, the evolved memory programs exhibit structurally distinct processing mechanisms for each domain. This finding indicates that specializing the memory mechanism for a given task explores a broad design space and provides a superior solution compared to general-purpose memory paradigms.", "published": "2026-04-10", "updated": "2026-05-23", "authors": [ "Wenbo Pan", "Shujie Liu", "Xiangyang Zhou", "Shiwei Zhang", "Wanlu Shi", "Mirror Xu", "Xiaohua Jia" ], "categories": [ "cs.PL", "cs.AI", "cs.CL", "cs.LG" ], "url": "https://arxiv.org/abs/2604.11811", "pdf_url": "https://arxiv.org/pdf/2604.11811", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2604.10981", "title": "ATANT v1.1: Positioning Continuity Evaluation Against Memory, Long-Context, and Agentic-Memory Benchmarks", "abstract": "ATANT v1.0 (arXiv:2604.06710) defined continuity as a system property with 7 required properties and introduced a 10-checkpoint, LLM-free evaluation methodology validated on a 250-story corpus. Since publication, a recurring reviewer and practitioner question has concerned not the framework itself but its relationship to a wider set of memory evaluations: LOCOMO, LongMemEval, BEAM, MemoryBench, Zep's evaluation suite, Letta/MemGPT's evaluations, and RULER. This companion paper, v1.1, does not modify the v1.0 standard. It closes a related-work gap that v1.0 left brief under page limits. We show by structural analysis that none of these benchmarks measures continuity as defined in v1.0: of the 7 required properties, the median existing eval covers 1 property, the mean covers 0.43 when partial credit is scored at 0.5, and no eval covers more than 2. We provide a cell-by-cell property-coverage matrix, identify methodological defects specific to each benchmark (including an empty-gold scoring bug in the LOCOMO reference implementation that renders 23% of its corpus unscorable by construction), and publish our reference implementation's LOCOMO score (8.8%) alongside the structural reason that number is uninformative about continuity. We publish our 8.8% LOCOMO score alongside our 96% ATANT cumulative-scale score as a calibration pair: the 87-point divergence is evidence that the two benchmarks measure different properties, not that one system is an order of magnitude better than another. The position v1.1 takes is not adversarial: each benchmark measures a real capability. The claim is that none of them can adjudicate continuity, and conflating them with continuity evaluation has led the field to under-invest in the properties v1.0 names.", "published": "2026-04-13", "updated": "2026-04-19", "authors": [ "Samuel Sameer Tanguturi" ], "categories": [ "cs.AI", "cs.IR" ], "url": "https://arxiv.org/abs/2604.10981", "pdf_url": "https://arxiv.org/pdf/2604.10981", "matched_queries": [ "agent-memory", "memory-evaluation" ] }, { "arxiv_id": "2604.11544", "title": "Time is Not a Label: Continuous Phase Rotation for Temporal Knowledge Graphs and Agentic Memory", "abstract": "Structured memory representations such as knowledge graphs are central to autonomous agents and other long-lived systems. However, most existing approaches model time as discrete metadata, either sorting by recency (burying old-yet-permanent knowledge), simply overwriting outdated facts, or requiring an expensive LLM call at every ingestion step, leaving them unable to distinguish persistent facts from evolving ones. To address this, we introduce RoMem, a drop-in temporal knowledge graph module for structured memory systems, applicable to agentic memory and beyond. A pretrained Semantic Speed Gate maps each relation's text embedding to a volatility score, learning from data that evolving relations (e.g., \"president of\") should rotate fast while persistent ones (e.g., \"born in\") should remain stable. Combined with continuous phase rotation, this enables geometric shadowing: obsolete facts are rotated out of phase in complex vector space, so temporally correct facts naturally outrank contradictions without deletion. On temporal knowledge graph completion, RoMem achieves state-of-the-art results on ICEWS05-15 (72.6 MRR). Applied to agentic memory, it delivers 2-3x MRR and answer accuracy on temporal reasoning (MultiTQ), dominates hybrid benchmark (LoCoMo), preserves static memory with zero degradation (DMR-MSC), and generalises zero-shot to unseen financial domains (FinTMMBench).", "published": "2026-04-13", "updated": "2026-04-13", "authors": [ "Weixian Waylon Li", "Jiaxin Zhang", "Xianan Jim Yang", "Tiejun Ma", "Yiwen Guo" ], "categories": [ "cs.CL", "cs.AI" ], "url": "https://arxiv.org/abs/2604.11544", "pdf_url": "https://arxiv.org/pdf/2604.11544", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2604.11563", "title": "Synthius-Mem: Brain-Inspired Hallucination-Resistant Persona Memory Achieving 94.4% Memory Accuracy and 99.6% Adversarial Robustness on LoCoMo", "abstract": "Providing AI agents with reliable long-term memory that does not hallucinate remains an open problem. Current approaches to memory for LLM agents -- sliding windows, summarization, embedding-based RAG, and flat fact extraction -- each reduce token cost but introduce catastrophic information loss, semantic drift, or uncontrolled hallucination about the user. The structural reason is architectural: every published memory system on the LoCoMo benchmark treats conversation as a retrieval problem over raw or lightly summarized dialogue segments, and none reports adversarial robustness, the ability to refuse questions about facts the user never disclosed. We present Synthius-Mem, a brain-inspired structured persona memory system that takes a fundamentally different approach. Instead of retrieving what was said, Synthius-Mem extracts what is known about the person: a full persona extraction pipeline decomposes conversations into six cognitive domains (biography, experiences, preferences, social circle, work, psychometrics), consolidates and deduplicates per domain, and retrieves structured facts via CategoryRAG at 21.79 ms latency. On the LoCoMo benchmark (ACL 2024, 10 conversations, 1,813 questions), Synthius-Mem achieves 94.37% accuracy, exceeding all published systems including MemMachine (91.69%, adversarial score is not reported) and human performance (87.9 F1). Core memory fact accuracy reaches 98.64%. Adversarial robustness, the hallucination resistance metric that no competing system reports, reaches 99.55%. Synthius-Mem reduces token consumption by ~5x compared to full-context replay while achieving higher accuracy. Synthius-Mem achieves state-of-the-art results on LoCoMo and is, to our knowledge, the only persona memory system that both exceeds human-level performance and reports adversarial robustness.", "published": "2026-04-13", "updated": "2026-04-13", "authors": [ "Artem Gadzhiev", "Andrew Kislov" ], "categories": [ "cs.CL", "cs.AI", "cs.LG" ], "url": "https://arxiv.org/abs/2604.11563", "pdf_url": "https://arxiv.org/pdf/2604.11563", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2604.12007", "title": "When to Forget: A Memory Governance Primitive", "abstract": "Agent memory systems accumulate experience but currently lack a principled operational metric for memory quality governance -- deciding which memories to trust, suppress, or deprecate as the agent's task distribution shifts. Write-time importance scores are static; dynamic management systems use LLM judgment or structural heuristics rather than outcome feedback. This paper proposes Memory Worth (MW): a two-counter per-memory signal that tracks how often a memory co-occurs with successful versus failed outcomes, providing a lightweight, theoretically grounded foundation for staleness detection, retrieval suppression, and deprecation decisions. We prove that MW converges almost surely to the conditional success probability p+(m) = Pr[y_t = +1 | m in M_t] -- the probability of task success given that memory m is retrieved -- under a stationary retrieval regime with a minimum exploration condition. Importantly, p+(m) is an associational quantity, not a causal one: it measures outcome co-occurrence rather than causal contribution. We argue this is still a useful operational signal for memory governance, and we validate it empirically in a controlled synthetic environment where ground-truth utility is known: after 10,000 episodes, the Spearman rank-correlation between Memory Worth and true utilities reaches rho = 0.89 +/- 0.02 across 20 independent seeds, compared to rho = 0.00 for systems that never update their assessments. A retrieval-realistic micro-experiment with real text and neural embedding retrieval (all-MiniLM-L6-v2) further shows stale memories crossing the low-value threshold (MW = 0.17) while specialist memories remain high-value (MW = 0.77) across 3,000 episodes. The estimator requires only two scalar counters per memory unit and can be added to architectures that already log retrievals and episode outcomes.", "published": "2026-04-13", "updated": "2026-04-13", "authors": [ "Baris Simsek" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2604.12007", "pdf_url": "https://arxiv.org/pdf/2604.12007", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2604.12179", "title": "AgenticAI-DialogGen: Topic-Guided Conversation Generation for Fine-Tuning and Evaluating Short- and Long-Term Memories of LLMs", "abstract": "Recent advancements in Large Language Models (LLMs) have improved their ability to process extended conversational contexts, yet fine-tuning and evaluating short- and long-term memories remain difficult due to the absence of datasets that encode both short- and long-term conversational history. Existing conversational datasets lack memory grounding, overlook topic continuity, or rely on costly human annotation. To address these gaps, we introduce AgenticAI-DialogGen, a modular agent-based framework that generates persona-grounded and topic-guided conversations without human supervision. The framework uses LLM agents to extract knowledge graphs, identify topics, build speaker personas, and simulate topic-guided conversations from unstructured conversations. A QA module generates memory-grounded Question Answer (QA) pairs drawn from short- and long-term conversational histories. We also generated a new dataset entitled, TopicGuidedChat (TGC), where long-term memory is encoded as speaker-specific knowledge graphs and short-term memory as newly generated topic-guided conversations. Evaluations depict that AgenticAI-DialogGen yields higher conversational quality and LLMs fine-tuned on TGC dataset achieve improved performance on memory-grounded QA tasks.", "published": "2026-04-14", "updated": "2026-04-14", "authors": [ "Manoj Madushanka Perera", "Adnan Mahmood", "Kasun Eranda Wijethilake", "Quan Z. Sheng" ], "categories": [ "cs.CL", "cs.IR" ], "url": "https://arxiv.org/abs/2604.12179", "pdf_url": "https://arxiv.org/pdf/2604.12179", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2604.12285", "title": "GAM: Hierarchical Graph-based Agentic Memory for LLM Agents", "abstract": "To sustain coherent long-term interactions, Large Language Model (LLM) agents must navigate the tension between acquiring new information and retaining prior knowledge. Current unified stream-based memory systems facilitate context updates but remain vulnerable to interference from transient noise. Conversely, discrete structured memory architectures provide robust knowledge retention but often struggle to adapt to evolving narratives. To address this, we propose GAM, a hierarchical Graph-based Agentic Memory framework that explicitly decouples memory encoding from consolidation to effectively resolve the conflict between rapid context perception and stable knowledge retention. By isolating ongoing dialogue in an event progression graph and integrating it into a topic associative network only upon semantic shifts, our approach minimizes interference while preserving long-term consistency. Additionally, we introduce a graph-guided, multi-factor retrieval strategy to enhance context precision. Experiments on LoCoMo and LongDialQA indicate that our method consistently outperforms state-of-the-art baselines in both reasoning accuracy and efficiency.", "published": "2026-04-14", "updated": "2026-04-14", "authors": [ "Zhaofen Wu", "Hanrong Zhang", "Fulin Lin", "Wujiang Xu", "Xinran Xu", "Yankai Chen", "Henry Peng Zou", "Shaowen Chen", "Weizhi Zhang", "Xue Liu", "Philip S. Yu", "Hongwei Wang" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2604.12285", "pdf_url": "https://arxiv.org/pdf/2604.12285", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2604.12948", "title": "Drawing on Memory: Dual-Trace Encoding Improves Cross-Session Recall in LLM Agents", "abstract": "LLM agents with persistent memory store information as flat factual records, providing little context for temporal reasoning, change tracking, or cross-session aggregation. Inspired by the drawing effect [3], we introduce dual-trace memory encoding. In this method, each stored fact is paired with a concrete scene trace, a narrative reconstruction of the moment and context in which the information was learned. The agent is forced to commit to specific contextual details during encoding, creating richer, more distinctive memory traces. Using the LongMemEval-S benchmark (4,575 sessions, 100 recall questions), we compare dual-trace encoding against a fact-only control with matched coverage and format over 99 shared questions. Dual-trace achieves 73.7% overall accuracy versus 53.5%, a +20.2 percentage point (pp) gain (95% CI: [+12.1, +29.3], bootstrap p < 0.0001). Gains concentrate in temporal reasoning (+40pp), knowledge-update tracking (+25pp), and multi-session aggregation (+30pp), with no benefit for single-session retrieval, consistent with encoding specificity theory [8]. Token analysis shows dual-trace encoding achieves this gain at no additional cost. We additionally sketch an architectural design for adapting dual-trace encoding to coding agents, with preliminary pilot validation.", "published": "2026-04-14", "updated": "2026-04-14", "authors": [ "Benjamin Stern", "Peter Nadel" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2604.12948", "pdf_url": "https://arxiv.org/pdf/2604.12948", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2604.15774", "title": "MemEvoBench: Benchmarking Safety Risks from Memory Misevolution in LLM Agents", "abstract": "Equipping Large Language Models (LLMs) with persistent memory enhances interaction continuity and personalization but introduces new safety risks. Specifically, contaminated or biased memory accumulation can trigger abnormal agent behaviors. Existing evaluation methods have not yet established a standardized framework for measuring memory misevolution. This phenomenon refers to the gradual behavioral drift resulting from repeated exposure to misleading information. To address this gap, we introduce MemEvoBench, the first benchmark evaluating long-horizon memory safety in LLM agents against adversarial memory injection, noisy tool outputs, and biased feedback. The framework consists of QA-style tasks across 7 domains and 36 risk types, complemented by workflow-style tasks adapted from 20 Agent-SafetyBench environments with noisy tool returns. Both settings employ mixed benign and misleading memory pools within multi-round interactions to simulate memory evolution. Experiments on representative models reveal substantial safety degradation under biased memory updates. Our analysis suggests that memory evolution is a significant contributor to these failures. Furthermore, static prompt-based defenses prove insufficient, underscoring the urgency of securing memory evolution in LLM agents.", "published": "2026-04-17", "updated": "2026-05-21", "authors": [ "Weiwei Xie", "Shaoxiong Guo", "Fan Zhang", "Tian Xia", "Xue Yang", "Lizhuang Ma", "Junchi Yan", "Qibing Ren" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2604.15774", "pdf_url": "https://arxiv.org/pdf/2604.15774", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2604.15877", "title": "Experience Compression Spectrum: Unifying Memory, Skills, and Rules in LLM Agents", "abstract": "As LLM agents scale to long-horizon, multi-session deployments, efficiently managing accumulated experience becomes a critical bottleneck. Agent memory systems and agent skill discovery both address this challenge, extracting reusable knowledge from interaction traces, yet a citation analysis of 1{,}136 references across 22 primary papers reveals a cross-community citation rate below 1\\%. We propose the \\emph{Experience Compression Spectrum}, a unifying framework that positions memory, skills, and rules as points along a single axis of increasing compression (5--20$\\times$ for episodic memory, 50--500$\\times$ for procedural skills, 1{,}000$\\times$+ for declarative rules), directly reducing context consumption, retrieval latency, and compute overhead. Mapping 20+ systems onto this spectrum reveals that every system operates at a fixed, predetermined compression level: none supports adaptive cross-level compression, a gap we term the \\emph{missing diagonal}. We further show that specialization alone is insufficient (both communities independently solve shared sub-problems without exchanging solutions), that evaluation methods are tightly coupled to compression levels, that transferability increases with compression at the cost of specificity, and that knowledge lifecycle management remains largely neglected. We articulate open problems and design principles for scalable, full-spectrum agent learning systems.", "published": "2026-04-17", "updated": "2026-06-25", "authors": [ "Xing Zhang", "Guanghui Wang", "Yanwei Cui", "Wei Qiu", "Ziyuan Li", "Bing Zhu", "Peiyang He" ], "categories": [ "cs.AI", "cs.CL", "cs.MA" ], "url": "https://arxiv.org/abs/2604.15877", "pdf_url": "https://arxiv.org/pdf/2604.15877", "matched_queries": [ "agent-memory", "episodic-memory" ] }, { "arxiv_id": "2604.16548", "title": "A Survey on Long-Term Memory Security in LLM Agents: Attacks, Defenses, and Governance Across the Memory Lifecycle", "abstract": "The emergence of writable, cross-session persistent memory in LLM agents introduces a qualitatively different threat landscape from conventional input-centric security concerns, characterized by three properties: persistence, statefulness, and propagation. To systematically characterize this landscape, we propose a Memory Lifecycle Framework that organizes attacks, defenses, and their cross-phase dependencies along two axes: six lifecycle phases (Write, Store, Retrieve, Execute, Share & Propagate, Forget & Rollback) and four security objectives (Integrity, Confidentiality, Availability, Governance). This analysis in turn exposes the need for formal security guarantees at the system level, motivating Verifiable Memory Governance(VMG), a framework of five architectural primitives that specifies what verifiable mechanisms a long-term-memory system must provide to maintain auditable, recoverable control over its memory state. Our analysis indicates that robust Long-Term Memory (LTM) security cannot be retrofitted at retrieval or execution time alone, but must be anchored in storage-time provenance, versioning, and policy-aware retention from the outset.", "published": "2026-04-17", "updated": "2026-06-11", "authors": [ "Zehao Lin", "Xixuan Hao", "Renyu Fu", "Shaobo Cui", "Kai Chen", "Chunyu Li", "Zhiyu Li", "Feiyu Xiong" ], "categories": [ "cs.CR", "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2604.16548", "pdf_url": "https://arxiv.org/pdf/2604.16548", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2604.16839", "title": "HeLa-Mem: Hebbian Learning and Associative Memory for LLM Agents", "abstract": "Long-term memory is a critical challenge for Large Language Model agents, as fixed context windows cannot preserve coherence across extended interactions. Existing memory systems represent conversation history as unstructured embedding vectors, retrieving information through semantic similarity. This paradigm fails to capture the associative structure of human memory, wherein related experiences progressively strengthen interconnections through repeated co-activation. Inspired by cognitive neuroscience, we identify three mechanisms central to biological memory: association, consolidation, and spreading activation, which remain largely absent in current research. To bridge this gap, we propose HeLa-Mem, a bio-inspired memory architecture that models memory as a dynamic graph with Hebbian learning dynamics. HeLa-Mem employs a dual-level organization: (1) an episodic memory graph that evolves through co-activation patterns, and (2) a semantic memory store populated via Hebbian Distillation, wherein a Reflective Agent identifies densely connected memory hubs and distills them into structured, reusable semantic knowledge. This dual-path design leverages both semantic similarity and learned associations, mirroring the episodic-semantic distinction in human cognition. Experiments on LoCoMo demonstrate superior performance across four question categories while using significantly fewer context tokens. Code is available on GitHub: https://github.com/ReinerBRO/HeLa-Mem", "published": "2026-04-18", "updated": "2026-04-18", "authors": [ "Jinchang Zhu", "Jindong Li", "Cheng Zhang", "Jiahong Liu", "Menglin Yang" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2604.16839", "pdf_url": "https://arxiv.org/pdf/2604.16839", "matched_queries": [ "episodic-memory", "long-term-memory" ] }, { "arxiv_id": "2604.17273", "title": "The Continuity Layer: Why Intelligence Needs an Architecture for What It Carries Forward", "abstract": "The most important architectural problem in AI is not the size of the model but the absence of a layer that carries forward what the model has come to understand. Sessions end. Context windows fill. Memory APIs return flat facts that the model has to reinterpret from scratch on every read. The result is intelligence that is powerful per session and amnesiac across time. This position paper argues that the layer which fixes this, the continuity layer, is the most consequential piece of infrastructure the field has not yet built, and that the engineering work to build it has begun in public. The formal evaluation framework for the property described here is the ATANT benchmark (arXiv:2604.06710), published separately with evaluation results on a 250-story corpus; a companion paper (arXiv:2604.10981) positions this framework against existing memory, long-context, and agentic-memory benchmarks. The paper defines continuity as a system property with seven required characteristics, distinct from memory and from retrieval; describes a storage primitive (Decomposed Trace Convergence Memory) whose write-time decomposition and read-time reconstruction produce that property; maps the engineering architecture to the theological pattern of kenosis and the symbolic pattern of Alpha and Omega, and argues this mapping is structural rather than metaphorical; proposes a four-layer development arc from external SDK to hardware node to long-horizon human infrastructure; examines why the physics limits now constraining the model layer make the continuity layer newly consequential; and argues that the governance architecture (privacy implemented as physics rather than policy, founder-controlled class shares on non-negotiable architectural commitments) is inseparable from the product itself.", "published": "2026-04-19", "updated": "2026-04-19", "authors": [ "Samuel Sameer Tanguturi" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2604.17273", "pdf_url": "https://arxiv.org/pdf/2604.17273", "matched_queries": [ "agent-memory", "memory-evaluation" ] }, { "arxiv_id": "2604.17456", "title": "TrafficClaw: A Generalizable LLM Agent in the Unified Physical Environment for Urban Traffic Control", "abstract": "Large language model (LLM) agents have shown strong capabilities in long-horizon reasoning, tool use, and decision-making in digital environments, yet extending them to physically grounded systems remains challenging. Unlike web, code, or game environments, where objectives are often weakly coupled, physical systems evolve through tightly coupled dynamics in which local interventions propagate across interacting subsystems over time. Urban traffic control exemplifies this challenge, as traffic signals, freeways, public transit, and taxi systems continuously interact through shared spatial infrastructure and temporal mobility demand. Existing optimization, reinforcement learning (RL), and LLM-based approaches are largely designed for isolated subsystems, limiting coordinated reasoning and system-level optimization. We propose TrafficClaw, a LLM-based generalizable traffic control agent for physical urban systems. TrafficClaw operates within a unified traffic environment that exposes coupled urban dynamics and feedback, performs executable spatiotemporal reasoning with persistent memory for long-horizon adaptation, and leverages multi-stage agentic RL for coordinated system-level optimization. Experiments across three metropolitan regions and six traffic-control tasks demonstrate strong generalization, robustness, and cross-subsystem coordination. Our project is available at https://github.com/usail-hkust/TrafficClaw.", "published": "2026-04-19", "updated": "2026-06-01", "authors": [ "Siqi Lai", "Pan Zhang", "Yuping Zhou", "Jindong Han", "Yansong Ning", "Hao Liu" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2604.17456", "pdf_url": "https://arxiv.org/pdf/2604.17456", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2604.17658", "title": "Towards Self-Improving Error Diagnosis in Multi-Agent Systems", "abstract": "Large Language Model (LLM)-based Multi-Agent Systems (MAS) enable complex problem-solving but introduce significant debugging challenges, characterized by long interaction traces, inter-agent dependencies, and delayed error manifestation. Existing diagnostic approaches often rely on expensive expert annotation or ''LLM-as-a-judge'' paradigms, which struggle to pinpoint decisive error steps within extended contexts. In this paper, we introduce ErrorProbe, a self-improving framework for semantic failure attribution that identifies responsible agents and the originating error step. The framework operates via a three-stage pipeline: (1) operationalizing the MAS failure taxonomy to detect local anomalies, (2) performing symptom-driven backward tracing to prune irrelevant context, and (3) employing a specialized multi-agent team (Strategist, Investigator, Arbiter) to validate error hypotheses through tool-grounded execution. Crucially, ErrorProbe maintains a verified episodic memory that updates only when error patterns are confirmed by executable evidence, without the need for annotation. Experiments across the TracerTraj and Who&When benchmarks demonstrate that ErrorProbe significantly outperforms baselines, particularly in step-level localization, while the verified memory enables robust cross-domain transfer without retraining.", "published": "2026-04-19", "updated": "2026-04-19", "authors": [ "Jiazheng Li", "Emine Yilmaz", "Bei Chen", "Dieu-Thu Le" ], "categories": [ "cs.MA", "cs.CL" ], "url": "https://arxiv.org/abs/2604.17658", "pdf_url": "https://arxiv.org/pdf/2604.17658", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2604.19541", "title": "FOCAL: Filtered On-device Continuous Activity Logging for Efficient Personal Desktop Summarization", "abstract": "Desktop interaction streams provide a continuous, privacy-sensitive record of interleaved user tasks. Transforming these streams into task-organized personal logs on-device faces two main challenges: exhaustive Vision-Language Model (VLM) processing strains local resources, and global stream processing causes cross-task context pollution. We present FOCAL (Filtered On-device Continuous Activity Logging), a privacy-first multi-agent system utilizing a unified filter-plan-log architecture. It cascades a lightweight Filter Agent for noise suppression, a text-only Brain Agent for task attribution, a Record Agent for selective visual reasoning, and a task-isolated Memory Agent for context-coherent summarization. Experiments on DesktopBench (comprising 2,572 screenshots across 420 complex sessions) show FOCAL reduces total token consumption by 60.4% and VLM call count by 72.3% versus a baseline, while boosting Key Information Recall (KIR) from 0.38 to 0.61. Crucially, under $A{\\to}B{\\to}A$ task interruptions, FOCAL maintains Task Acc 0.81 and KIR 0.80, whereas the baseline collapses to Task Acc 0.03. FOCAL pioneers the efficient, on-device summarization of instruction-free desktop streams into multi-perspective personal logs.", "published": "2026-04-21", "updated": "2026-04-21", "authors": [ "Haoran Yin", "Zhiyuan Wen", "Jiannong Cao", "Bo Yuan", "Ruosong Yang" ], "categories": [ "cs.MA", "cs.HC" ], "url": "https://arxiv.org/abs/2604.19541", "pdf_url": "https://arxiv.org/pdf/2604.19541", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2604.20006", "title": "From Recall to Forgetting: Benchmarking Long-Term Memory for Personalized Agents", "abstract": "Personalized agents that interact with users over long periods must maintain persistent memory across sessions and update it as circumstances change. However, existing benchmarks predominantly frame long-term memory evaluation as fact retrieval from past conversations, providing limited insight into agents' ability to consolidate memory over time or handle frequent knowledge updates. We introduce Memora, a long-term memory benchmark spanning weeks to months long user conversations. The benchmark evaluates three memory-grounded tasks: remembering, reasoning, and recommending. To ensure data quality, we employ automated memory-grounding checks and human evaluation. We further introduce Forgetting-Aware Memory Accuracy (FAMA), a metric that penalizes reliance on obsolete or invalidated memory when evaluating long-term memory. Evaluations of four LLMs and six memory agents reveal frequent reuse of invalid memories and failures to reconcile evolving memories. Memory agents offer marginal improvements, exposing shortcomings in long-term memory for personalized agents.", "published": "2026-04-21", "updated": "2026-04-21", "authors": [ "Md Nayem Uddin", "Kumar Shubham", "Eduardo Blanco", "Chitta Baral", "Gengyu Wang" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2604.20006", "pdf_url": "https://arxiv.org/pdf/2604.20006", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2604.20117", "title": "To Know is to Construct: Schema-Constrained Generation for Agent Memory", "abstract": "Constructivist epistemology argues that knowledge is actively constructed rather than passively copied. Despite the generative nature of Large Language Models (LLMs), most existing agent memory systems are still based on dense retrieval. However, dense retrieval heavily relies on semantic overlap or entity matching within sentences. Consequently, embeddings often fail to distinguish instances that are semantically similar but contextually distinct, introducing substantial noise by retrieving context-mismatched entries. Conversely, directly employing open-ended generation for memory access risks \"Structural Hallucination\" where the model generates memory keys that do not exist in the memory, leading to lookup failures. Inspired by this epistemology, we posit that memory is fundamentally organized by cognitive schemas, and valid recall must be a generative process performed within these schematic structures. To realize this, we propose SCG-MEM, a schema-constrained generative memory architecture. SCG-MEM reformulates memory access as Schema-Constrained Generation. By maintaining a dynamic Cognitive Schema, we strictly constrain LLM decoding to generate only valid memory entry keys, providing a formal guarantee against structural hallucinations. To support long-term adaptation, we model memory updates via assimilation (grounding inputs into existing schemas) and accommodation (expanding schemas with novel concepts). Furthermore, we construct an Associative Graph to enable multi-hop reasoning through activation propagation. Experiments on the LoCoMo benchmark show that SCG-MEM substantially improves performance across all categories over retrieval-based baselines.", "published": "2026-04-22", "updated": "2026-04-22", "authors": [ "Lei Zheng", "Weinan Song", "Daili Li", "Yanming Yang" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2604.20117", "pdf_url": "https://arxiv.org/pdf/2604.20117", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2604.20183", "title": "Dual-Cluster Memory Agent: Resolving Multi-Paradigm Ambiguity in Optimization Problem Solving", "abstract": "Large Language Models (LLMs) often struggle with structural ambiguity in optimization problems, where a single problem admits multiple related but conflicting modeling paradigms, hindering effective solution generation. To address this, we propose Dual-Cluster Memory Agent (DCM-Agent) to enhance performance by leveraging historical solutions in a training-free manner. Central to this is Dual-Cluster Memory Construction. This agent assigns historical solutions to modeling and coding clusters, then distills each cluster's content into three structured types: Approach, Checklist, and Pitfall. This process derives generalizable guidance knowledge. Furthermore, this agent introduces Memory-augmented Inference to dynamically navigate solution paths, detect and repair errors, and adaptively switch reasoning paths with structured knowledge. The experiments across seven optimization benchmarks demonstrate that DCM-Agent achieves an average performance improvement of 11%- 21%. Notably, our analysis reveals a ``knowledge inheritance'' phenomenon: memory constructed by larger models can guide smaller models toward superior performance, highlighting the framework's scalability and efficiency.", "published": "2026-04-22", "updated": "2026-06-02", "authors": [ "Xinyu Zhang", "Yuchen Wan", "Boxuan Zhang", "Zesheng Yang", "Lingling Zhang", "Bifan Wei", "Jun Liu" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2604.20183", "pdf_url": "https://arxiv.org/pdf/2604.20183", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2604.20300", "title": "FSFM: A Biologically-Inspired Framework for Selective Forgetting of Agent Memory", "abstract": "For LLM agents, memory management critically impacts efficiency, quality, and security. While much research focuses on retention, selective forgetting--inspired by human cognitive processes (hippocampal indexing/consolidation theory and Ebbinghaus forgetting curve)--remains underexplored. We argue that in resource-constrained environments, a well-designed forgetting mechanism is as crucial as remembering, delivering benefits across three dimensions: (1) efficiency via intelligent memory pruning, (2) quality by dynamically updating outdated preferences and context, and (3) security through active forgetting of malicious inputs, sensitive data, and privacy-compromising content. Our framework establishes a taxonomy of forgetting mechanisms: passive decay-based, active deletion-based, safety-triggered, and adaptive reinforcement-based. Building on advances in LLM agent architectures and vector databases, we present detailed specifications, implementation strategies, and empirical validation from controlled experiments. Results show significant improvements: access efficiency (+8.49%), content quality (+29.2% signal-to-noise ratio), and security performance (100% elimination of security risks). Our work bridges cognitive neuroscience and AI systems, offering practical solutions for real-world deployment while addressing ethical and regulatory compliance. The paper concludes with challenges and future directions, establishing selective forgetting as a fundamental capability for next-generation LLM agents operating in real-world, resource-constrained scenarios. Our contributions align with AI-native memory systems and responsible AI development.", "published": "2026-04-22", "updated": "2026-04-23", "authors": [ "Yingjie Gu", "Wenjian Xiong", "Liqiang Wang", "Pengcheng Ren", "Chao Li", "Xiaojing Zhang", "Yijuan Guo", "Qi Sun", "Jingyao Ma", "Shidang Shi" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2604.20300", "pdf_url": "https://arxiv.org/pdf/2604.20300", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2604.20582", "title": "Trust, Lies, and Long Memories: Emergent Social Dynamics and Reputation in Multi-Round Avalon with LLM Agents", "abstract": "We study emergent social dynamics in LLM agents playing The Resistance: Avalon, a hidden-role deception game. Unlike prior work on single-game performance, our agents play repeated games while retaining memory of previous interactions, including who played which roles and how they behaved, enabling us to study how social dynamics evolve. Across 188 games, two key phenomena emerge. First, reputation dynamics emerge organically when agents retain cross-game memory: agents reference past behavior in statements like \"I am wary of repeating last game's mistake of over-trusting early success.\" These reputations are role-conditional: the same agent is described as \"straightforward\" when playing good but \"subtle\" when playing evil, and high-reputation players receive 46% more team inclusions. Second, higher reasoning effort supports more strategic deception: evil players more often pass early missions to build trust before sabotaging later ones, 75% in high-effort games vs 36% in low-effort games. Together, these findings show that repeated interaction with memory gives rise to measurable reputation and deception dynamics among LLM agents.", "published": "2026-04-22", "updated": "2026-04-22", "authors": [ "Suveen Ellawela" ], "categories": [ "cs.MA", "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2604.20582", "pdf_url": "https://arxiv.org/pdf/2604.20582", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2604.20598", "title": "Self-Aware Vector Embeddings for Retrieval-Augmented Generation: A Neuroscience-Inspired Framework for Temporal, Confidence-Weighted, and Relational Knowledge", "abstract": "Modern retrieval-augmented generation (RAG) systems treat vector embeddings as static, context-free artifacts: an embedding has no notion of when it was created, how trustworthy its source is, or which other embeddings depend on it. This flattening of knowledge has a measurable cost: recent work on VersionRAG reports that conventional RAG achieves only 58% accuracy on versioned technical queries, because retrieval returns semantically similar but temporally invalid content. We propose SmartVector, a framework that augments dense embeddings with three explicit properties -- temporal awareness, confidence decay, and relational awareness -- and a five-stage lifecycle modeled on hippocampal-neocortical memory consolidation. A retrieval pipeline replaces pure cosine similarity with a four-signal score that mixes semantic relevance, temporal validity, live confidence, and graph-relational importance. A background consolidation agent detects contradictions, builds dependency edges, and propagates updates along those edges as graph-neural-network-style messages. Confidence is governed by a closed-form function combining an Ebbinghaus-style exponential decay, user-feedback reconsolidation, and logarithmic access reinforcement. We formalize the model, relate it to temporal knowledge graph embedding, agentic memory architectures, and uncertainty-aware RAG, and present a reference implementation. On a reproducible synthetic versioned-policy benchmark of 258 vectors and 138 queries, SmartVector roughly doubles top-1 accuracy over plain cosine RAG (62.0% vs. 31.0% on a held-out split), drops stale-answer rate from 35.0% to 13.3%, cuts Expected Calibration Error by nearly 2x (0.244 vs. 0.470), reduces re-embedding cost per single-word edit by 77%, and is robust across contradiction-injection rates from 0% to 75%.", "published": "2026-04-22", "updated": "2026-04-22", "authors": [ "Naizhong Xu" ], "categories": [ "cs.IR", "cs.CL", "cs.DB", "cs.LG" ], "url": "https://arxiv.org/abs/2604.20598", "pdf_url": "https://arxiv.org/pdf/2604.20598", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2604.22085", "title": "Memanto: Typed Semantic Memory with Information-Theoretic Retrieval for Long-Horizon Agents", "abstract": "The transition from stateless language model inference to persistent, multi session autonomous agents has revealed memory to be a primary architectural bottleneck in the deployment of production grade agentic systems. Existing methodologies largely depend on hybrid semantic graph architectures, which impose substantial computational overhead during both ingestion and retrieval. These systems typically require large language model mediated entity extraction, explicit graph schema maintenance, and multi query retrieval pipelines. This paper introduces Memanto, a universal memory layer for agentic artificial intelligence that challenges the prevailing assumption that knowledge graph complexity is necessary to achieve high fidelity agent memory. Memanto integrates a typed semantic memory schema comprising thirteen predefined memory categories, an automated conflict resolution mechanism, and temporal versioning. These components are enabled by Moorcheh's Information Theoretic Search engine, a no indexing semantic database that provides deterministic retrieval within sub ninety millisecond latency while eliminating ingestion delay. Through systematic benchmarking on the LongMemEval and LoCoMo evaluation suites, Memanto achieves state of the art accuracy scores of 89.8 percent and 87.1 percent respectively. These results surpass all evaluated hybrid graph and vector based systems while requiring only a single retrieval query, incurring no ingestion cost, and maintaining substantially lower operational complexity. A five stage progressive ablation study is presented to quantify the contribution of each architectural component, followed by a discussion of the implications for scalable deployment of agentic memory systems.", "published": "2026-04-23", "updated": "2026-04-23", "authors": [ "Seyed Moein Abtahi", "Rasa Rahnema", "Hetkumar Patel", "Neel Patel", "Majid Fekri", "Tara Khani" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2604.22085", "pdf_url": "https://arxiv.org/pdf/2604.22085", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2604.23711", "title": "Spore: Efficient and Training-Free Privacy Extraction Attack on LLMs via Inference-Time Hybrid Probing", "abstract": "With the wide adoption of personal AI assistants such as OpenClaw, privacy leakage in user interaction contexts with large language model (LLM) agents has become a critical issue. Existing privacy attacks against LLMs primarily target training data, while research on inference-time contextual privacy risks in LLM agent memory remains limited. Moreover, prior methods often incur high attack costs, requiring multiple queries or relying on white-box assumptions, which limits their practicality in real-world deployments. To address these issues, we propose a training-free privacy extraction attack targeting LLM agent memory, which we name \\textsc{Spore}. \\textsc{Spore} is compatible with both black-box and gray-box settings. In the black-box setting, \\textsc{Spore} can efficiently extract a small candidate set via a single query to recover the original private information. In the gray-box setting, \\textsc{Spore} allows the attacker to leverage multi-ranked tokens for more accurate and faster privacy extraction. We provide an information-theoretic analysis of \\textsc{Spore} and show that it achieves high query efficiency with substantial per query information leakage. Experiments on multiple frontier LLMs show that \\textsc{Spore} outperforms attack success rate over existing state-of-the-art (SOTA) schemes. It also maintains low attack cost and remains stable across different model parameter settings. We further evaluate the robustness of \\textsc{Spore} against existing defense mechanisms. Our results show that \\textsc{Spore} consistently bypasses both detection and strong safety alignment, demonstrating resilient performance in diverse defensive settings and real-world safety threats.", "published": "2026-04-26", "updated": "2026-04-26", "authors": [ "Yu Cui", "Ruiqing Yue", "Hang Fu", "Sicheng Pan", "Zhuoyu Sun", "Baohan Huang", "Haibin Zhang", "Cong Zuo", "Licheng Wang" ], "categories": [ "cs.CR" ], "url": "https://arxiv.org/abs/2604.23711", "pdf_url": "https://arxiv.org/pdf/2604.23711", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.20570", "title": "Infrastructure for the Agentic Web: Gap Analysis and Architecture from the Agentverse Platform", "abstract": "The emergence of autonomous AI agents as first-class participants in digital infrastructure marks a fundamental inflection point in the evolution of the Web. While significant research has been directed at agent behaviour and reasoning, comparatively little attention has been paid to the infrastructure those agents require to operate reliably at scale. This paper addresses that gap with a systematic analysis of Agentverse, the agent cloud platform developed by Fetch.ai under the Artificial Superintelligence (ASI) Alliance, which represents one of the most mature production deployments of agent-native infrastructure available today. We make three principal contributions. First, we conduct an empirical audit of the Agentverse platform, cataloguing 204 API endpoints (Q1 2026) and characterising what is operational, partially deployed, or absent. From this audit we derive a Gap Taxonomy of eight categories encompassing 62 distinct missing capabilities, ranging from agent memory and observability to security, economic primitives, and enterprise scaling. Second, we propose a seven-layer Agent Cloud Stack -- a reference architecture for what a fully realised agent-native cloud should provide by 2030, grounded in the specific gaps we identify. Third, we characterise five critical evolution paths: from ephemeral storage to a full Agent Memory Cloud; from keyword discovery to a semantic, trust-weighted Agent DNS; from a single-protocol model to a multi-standard agent lingua franca; from single-instance hosting to Kubernetes-scale orchestration; and from simple token payments to rich agent economic primitives. Together these contributions provide a diagnostic of current agent infrastructure and a technically grounded vision for what the agent cloud must become to support the agentic web -- Web4 -- by 2030.", "published": "2026-04-26", "updated": "2026-04-26", "authors": [ "Robin Dey", "Panyanon Viradecha" ], "categories": [ "cs.NI", "cs.AI", "cs.DC", "cs.MA" ], "url": "https://arxiv.org/abs/2606.20570", "pdf_url": "https://arxiv.org/pdf/2606.20570", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2604.26197", "title": "Hierarchical Long-Term Semantic Memory for LinkedIn's Hiring Agent", "abstract": "Large Language Model (LLM) agents are increasingly used in real-world products, where personalized and context-aware user interactions are essential. A central enabler of such capabilities is the agent's long-term semantic memory system, which extracts implicit and explicit signals from noisy longitudinal behavioral data, stores them in a structured form, and supports low-latency retrieval. Building industrial-grade long-term memory for LLM agents raises five challenges: scalability, low-latency retrieval, privacy constraints, adaptability, and observability. We introduce the Hierarchical Long-Term Semantic Memory (HLTM) framework, which organizes textual data into a schema-aligned memory tree that captures semantic knowledge at multiple levels of granularity, enabling scalable ingestion, privacy-aware storage, low-latency retrieval, and transparent provenance; HLTM further incorporates an adaptation mechanism to generalize across diverse use cases. Extensive evaluations on LinkedIn's Hiring Assistant show that HLTM improves answer correctness by more than 5% and retrieval F1 by more than 10%, while significantly advancing the Pareto frontier between query and indexing latency. HLTM has been fully deployed in LinkedIn's Hiring Assistant to power core personalization features in production hiring workflows.", "published": "2026-04-29", "updated": "2026-05-30", "authors": [ "Zhentao Xu", "Shangjin Zhang", "Emir Poyraz", "Yvonne Li", "Ye Jin", "Xie Lu", "Xiaoyang Gu", "Karthik Ramgopal", "Praveen Kumar Bodigutla", "Xiaofeng Wang" ], "categories": [ "cs.IR", "cs.LG" ], "url": "https://arxiv.org/abs/2604.26197", "pdf_url": "https://arxiv.org/pdf/2604.26197", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2604.26622", "title": "OCR-Memory: Optical Context Retrieval for Long-Horizon Agent Memory", "abstract": "Autonomous LLM agents increasingly operate in long-horizon, interactive settings where success depends on reusing experience accumulated over extended histories. However, existing agent memory systems are fundamentally constrained by text-context budgets: storing or revisiting raw trajectories is prohibitively token-expensive, while summarization and text-only retrieval trade token savings for information loss and fragmented evidence. To address this limitation, we propose Optical Context Retrieval Memory (OCR-Memory), a memory framework that leverages the visual modality as a high-density representation of agent experience, enabling retention of arbitrarily long histories with minimal prompt overhead at retrieval time. Specifically, OCR-Memory renders historical trajectories into images annotated with unique visual identifiers. OCR-Memory retrieves stored experience via a \\emph{locate-and-transcribe} paradigm that selects relevant regions through visual anchors and retrieves the corresponding verbatim text, avoiding free-form generation and reducing hallucination. Experiments on long-horizon agent benchmarks show consistent gains under strict context limits, demonstrating that optical encoding increases effective memory capacity while preserving faithful evidence recovery.", "published": "2026-04-29", "updated": "2026-04-29", "authors": [ "Jinze Li", "Yang Zhang", "Xin Yang", "Jiayi Qu", "Jinfeng Xu", "Shuo Yang", "Junhua Ding", "Edith Cheuk-Han Ngai" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2604.26622", "pdf_url": "https://arxiv.org/pdf/2604.26622", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2604.27003", "title": "When Continual Learning Moves to Memory: A Study of Experience Reuse in LLM Agents", "abstract": "Memory-augmented LLM agents offer an appealing shortcut to continual learning: rather than updating model parameters, they accumulate experience in external memory, seemingly sidestepping the stability-plasticity dilemma of parametric learning. We show that this challenge does not disappear but resurfaces at the memory level. Under a limited context window, old and new experiences compete during retrieval, relocating the continual-learning bottleneck from parameter updates to memory access. To study this phenomenon, we introduce a (k,v) framework that disentangles two fundamental design axes of external memory: how experience is represented and how it is organized for retrieval. Across sequential-task experiments in ALFWorld and BabyAI, we find that abstract procedural memories transfer more reliably than detailed trajectories, while negative transfer disproportionately harms the hard cases. Moreover, finer-grained memory organization is not universally beneficial: designs that yield strong forward transfer can simultaneously induce severe forgetting. Together, these results reveal that external memory does not resolve the continual-learning problem; it reshapes it into a problem of memory representation and retrieval design.", "published": "2026-04-29", "updated": "2026-04-29", "authors": [ "Qisheng Hu", "Quanyu Long", "Wenya Wang" ], "categories": [ "cs.LG", "cs.AI" ], "url": "https://arxiv.org/abs/2604.27003", "pdf_url": "https://arxiv.org/pdf/2604.27003", "matched_queries": [ "procedural-memory" ] }, { "arxiv_id": "2604.27045", "title": "Detecting Clinical Discrepancies in Health Coaching Agents: A Dual-Stream Memory and Reconciliation Architecture", "abstract": "As Large Language Model (LLM) agents transition from single-session tools to persistent systems managing longitudinal healthcare journeys, their memory architectures face a critical challenge: reconciling two imperfect sources of truth. The patient's evolving self-report is current but prone to recall bias, while the Electronic Health Record (EHR) is medically validated but frequently stale. General-purpose agent memory systems optimize for coherence by overwriting older facts with the user's latest statement, a pattern that risks safety failures when applied to clinical data. We introduce a Dual-Stream Memory Architecture that strictly separates the patient narrative from the structured clinical record (FHIR), governed by a dedicated Reconciliation Engine that evaluates every extracted memory against the patient's FHIR profile and classifies discrepancies by type, severity, and the specific FHIR resources involved. We evaluate this architecture on 26 patients across 675 longitudinal wellness coaching sessions, using a hybrid dataset that interleaves real provider-patient transcripts with synthetic, FHIR-grounded clinical scenarios. In isolated testing, the engine detects 84.4% of designed clinical discrepancies with 86.7% safety-critical recall. By coupling extraction and reconciliation evaluation on the same data, we directly quantify a 13.6% error cascade, tracing the degradation to clinical details lost during memory extraction from unstructured conversation rather than to downstream classification errors. These findings establish that validating patient-reported memories against clinical records is both feasible and necessary for safe deployment of longitudinal health agents.", "published": "2026-04-29", "updated": "2026-04-29", "authors": [ "Samuel L Pugh", "Eric Yang", "Alexander Muir Sutherland", "Alessandra Breschi" ], "categories": [ "cs.LG", "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2604.27045", "pdf_url": "https://arxiv.org/pdf/2604.27045", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2604.27283", "title": "Learning When to Remember: Risk-Sensitive Contextual Bandits for Abstention-Aware Memory Retrieval in LLM-Based Coding Agents", "abstract": "Large language model (LLM)-based coding agents increasingly rely on external memory to reuse prior debugging experience, repair traces, and repository-local operational knowledge. However, retrieved memory is useful only when the current failure is genuinely compatible with a previous one; superficial similarity in stack traces, terminal errors, paths, or configuration symptoms can lead to unsafe memory injection. This paper reframes issue-memory use as a selective, risk-sensitive control problem rather than a pure top-k retrieval problem. We introduce RSCB-MC, a risk-sensitive contextual bandit memory controller that decides whether an agent should use no memory, inject the top resolution, summarize multiple candidates, perform high-precision or high-recall retrieval, abstain, or ask for feedback. The system stores reusable issue knowledge through a pattern-variant-episode schema and converts retrieval evidence into a fixed 16-feature contextual state capturing relevance, uncertainty, structural compatibility, feedback history, false-positive risk, latency, and token cost. Its reward design penalizes false-positive memory injection more strongly than missed reuse, making non-injection and abstention first-class safety actions. In deterministic smoke-scale artifacts, RSCB-MC obtains the strongest non-oracle offline replay success rate, 62.5%, while maintaining a 0.0% false-positive rate. In a bounded 200-case hot-path validation, it reaches 60.5% proxy success with 0.0% false positives and a 331.466 microseconds p95 decision latency. The results show that, for coding-agent memory, the key question is not only which memory is most similar, but whether any retrieved memory is safe enough to influence the debugging trajectory.", "published": "2026-04-30", "updated": "2026-04-30", "authors": [ "Mehmet Iscan" ], "categories": [ "cs.CL", "cs.AI", "cs.LG" ], "url": "https://arxiv.org/abs/2604.27283", "pdf_url": "https://arxiv.org/pdf/2604.27283", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2604.27707", "title": "Contextual Agentic Memory is a Memo, Not True Memory", "abstract": "Current agentic memory systems (vector stores, retrieval-augmented generation, scratchpads, and context-window management) do not implement memory: they implement lookup. We argue that treating lookup as memory is a category error with provable consequences for agent capability, long-term learning, and security. Retrieval generalizes by similarity to stored cases; weight-based memory generalizes by applying abstract rules to inputs never seen before. Conflating the two produces agents that accumulate notes indefinitely without developing expertise, face a provable generalization ceiling on compositionally novel tasks that no increase in context size or retrieval quality can overcome, and are structurally vulnerable to persistent memory poisoning as injected content propagates across all future sessions. Drawing on Complementary Learning Systems theory from neuroscience, we show that biological intelligence solved this problem by pairing fast hippocampal exemplar storage with slow neocortical weight consolidation, and that current AI agents implement only the first half. We formalize these limitations, address four alternative views, and close with a co-existence proposal and a call to action for system builders, benchmark designers, and the memory community.", "published": "2026-04-30", "updated": "2026-04-30", "authors": [ "Binyan Xu", "Xilin Dai", "Kehuan Zhang" ], "categories": [ "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2604.27707", "pdf_url": "https://arxiv.org/pdf/2604.27707", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2604.27996", "title": "Exploring LLM Agent Designs and Interaction Modalities for Scientific Visualization", "abstract": "This paper examines how large language model (LLM) agents perform on scientific visualization (SciVis) tasks that require generating visualization workflows from natural-language instructions. We compare three representative agent designs: domain-specific agents with structured tool use, computer-use agents, and general-purpose coding agents, across 15 benchmark tasks, evaluating visualization quality, efficiency, robustness, computational cost, and the impact of persistent memory. We further study interaction modalities, including code scripts, model context protocol (MCP) or API calls, command-line interfaces (CLI), and graphical user interfaces (GUI). Our goal is to characterize the tradeoffs among representative SciVis agent configurations used in practice. The results reveal clear tradeoffs across agent designs and interaction modalities. General-purpose coding agents achieve the highest task success rates but incur greater computational cost, whereas domain-specific agents are more efficient and stable but less flexible. Computer-use agents perform well on individual operations but struggle with multi-step workflows. Across both CLI- and GUI-based settings, persistent memory improves performance over repeated trials, but its effectiveness depends on the interaction mode and the quality of feedback. These findings suggest that future SciVis systems should combine structured tool use, interactive capabilities, and adaptive memory mechanisms to balance performance, robustness, and flexibility.", "published": "2026-04-30", "updated": "2026-06-26", "authors": [ "Jackson Vonderhorst", "Kuangshi Ai", "Haichao Miao", "Shusen Liu", "Chaoli Wang" ], "categories": [ "cs.AI", "cs.GR", "cs.HC" ], "url": "https://arxiv.org/abs/2604.27996", "pdf_url": "https://arxiv.org/pdf/2604.27996", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2605.00702", "title": "Learning How and What to Memorize: Cognition-Inspired Two-Stage Optimization for Evolving Memory", "abstract": "Large language model (LLM) agents require long-term user memory for consistent personalization, but limited context windows hinder tracking evolving preferences over long interactions. Existing memory systems mainly rely on static, hand-crafted update rules; although reinforcement learning (RL)-based agents learn memory updates, sparse outcome rewards provide weak supervision, resulting in unstable long-horizon optimization. Drawing on memory schema theory and the functional division between prefrontal regions and hippocampus regions, we introduce MemCoE, a cognition-inspired two-stage optimization framework that learns how memory should be organized and what information to update. In the first stage, we propose Memory Guideline Induction to optimize a global guideline via contrastive feedback interpreted as textual gradients; in the second stage, Guideline-Aligned Memory Policy Optimization uses the induced guideline to define structured process rewards and performs multi-turn RL to learn a guideline-following memory evolution policy. We evaluate on three personalization memory benchmarks, covering explicit/implicit preference and different sizes and noise, and observe consistent improvements over strong baselines with favorable robustness, transferability, and efficiency.", "published": "2026-05-01", "updated": "2026-05-01", "authors": [ "Derong Xu", "Shuochen Liu", "Pengfei Luo", "Pengyue Jia", "Yingyi Zhang", "Yi Wen", "Yimin Deng", "Wenlin Zhang", "Enhong Chen", "Xiangyu Zhao", "Tong Xu" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2605.00702", "pdf_url": "https://arxiv.org/pdf/2605.00702", "matched_queries": [ "memory-evaluation" ] }, { "arxiv_id": "2605.01386", "title": "MemORAI: Memory Organization and Retrieval via Adaptive Graph Intelligence for LLM Conversational Agents", "abstract": "Large Language Models (LLMs) lack persistent memory for long-term personalized conversations. Existing graph-based memory systems suffer from information dilution, absent provenance tracking, and uniform retrieval that ignores query context. We introduce MemORAI (Memory Organization and Retrieval via Adaptive Graph Intelligence), a framework that integrates three innovations: selective memory filtering with dual-layer compression to retain user-persona-relevant content, a provenance-enriched multi-relational graph tracking factual origins at the turn level, and query-adaptive subgraph retrieval with Dynamic Weighted PageRank that applies query-conditioned edge weighting. Evaluated on LOCOMO and LongMemEval benchmarks, MemORAI achieves state-of-the-art performance in memory retrieval and personalized response generation, demonstrating that selective storage, enriched representation, and adaptive retrieval are essential for coherent, personalized LLM agents.", "published": "2026-05-02", "updated": "2026-06-02", "authors": [ "Hung Pham Van", "Nguyen Manh Hieu", "Khang Pham Tran Tuan", "Nam Le Hai", "Linh Ngo Van", "Nguyen Thi Ngoc Diep", "Trung Le" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2605.01386", "pdf_url": "https://arxiv.org/pdf/2605.01386", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2605.01970", "title": "Trojan Hippo: Weaponizing Agent Memory for Data Exfiltration", "abstract": "Memory systems enable otherwise-stateless LLM agents to persist user information across sessions, but also introduce a new attack surface. We characterize the Trojan Hippo attack, a class of persistent memory attacks that operates in a more realistic threat model than prior memory poisoning work: the attacker plants a dormant payload into an agent's long-term memory via a single untrusted tool call (e.g., a crafted email), which activates only when the user later discusses sensitive topics such as finance, health, or identity, and exfiltrates high-value personal data to the attacker. While anecdotal demonstrations of such attacks have appeared against deployed systems, no prior work systematically evaluates them across heterogeneous memory architectures and defenses. We introduce a dynamic evaluation framework comprising two components: (1) an OpenEvolve-based adaptive red-teaming benchmark that stress-tests defenses and memory backends against continuously refined attacks, and (2) the first capability-aware security/utility analysis for persistent memory systems, enabling principled reasoning about defense deployment across different usage profiles. Instantiated on an email assistant across four memory backends (explicit tool memory, agentic memory, RAG, and sliding-window context), Trojan Hippo achieves up to 85-100% ASR against current frontier models from OpenAI and Google, with planted memories successfully activating even after 100 benign sessions. We evaluate four memory-system defenses inspired by basic security principles, finding they substantially reduce attack success rates (to as low as 0-5%), though at utility costs that vary widely with task requirements. Because of this substantial security-utility tradeoff, the effective real-world deployment of defenses remains an open challenge, which our evaluation framework is specifically designed to address.", "published": "2026-05-03", "updated": "2026-05-15", "authors": [ "Debeshee Das", "Julien Piet", "Darya Kaviani", "Luca Beurer-Kellner", "Florian Tramèr", "David Wagner" ], "categories": [ "cs.CR", "cs.AI" ], "url": "https://arxiv.org/abs/2605.01970", "pdf_url": "https://arxiv.org/pdf/2605.01970", "matched_queries": [ "agent-memory", "long-term-memory" ] }, { "arxiv_id": "2605.02199", "title": "MEMAUDIT: An Exact Package-Oracle Evaluation Protocol for Budgeted Long-Term LLM Memory Writing", "abstract": "Long-term LLM agents must compress streams of past interactions into persistent memory before future queries are known. Existing evaluations usually measure final question-answering accuracy, which entangles memory writing with retrieval, prompting, and reader reasoning. We introduce MEMAUDIT, an exact packageoracle evaluation protocol for budgeted long-term memory writing. A MEMAUDIT package fixes an experience stream, candidate memory representations, storage costs, semantic evidence units, future-query requirements, and a budget, turning write-time memory selection into a finite auditable optimization problem with a certified denominator. We instantiate this protocol with a concave-over-modular semantic coverage objective under storage and one-representation-per-experience constraints, and compute exact package optima using branch-and-bound with MILP certification. Across controlled exact packages, validity-heavy stress tests, human-audited natural support slices, and exported Mem0, A-Mem, and Letta stores, MEMAUDIT separates representation quality, validity-state preservation, and budget-aware selection effects that end-to-end QA cannot localize. The resulting artifact provides reusable package generators, certified solvers, natural package exports, external-system scorers, and cached reproducibility metadata for evaluating what memory writers actually preserve under fixed storage budgets.", "published": "2026-05-04", "updated": "2026-05-04", "authors": [ "Nishant Bhargava", "Rodrigo Sobral Barrento" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2605.02199", "pdf_url": "https://arxiv.org/pdf/2605.02199", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2605.03228", "title": "MAGE: Safeguarding LLM Agents against Long-Horizon Threats via Shadow Memory", "abstract": "As large language model (LLM)-powered agents are increasingly deployed to perform complex, real-world tasks, they face a growing class of attacks that exploit extended user-agent-environment interactions to pursue malicious objectives improbable in single-turn settings. Such long-horizon threats pose significant risks to the safe deployment of LLM agents in critical domains. In this paper, we present MAGE (Memory As Guardrail Enforcement), a novel defensive framework designed to counter a wide range of long-horizon threats. Inspired by the \"shadow stack\" abstraction in systems security, MAGE maintains a dedicated, safety-focused agentic memory that distills and retains safety-critical context across the agent's full execution trajectory, leveraging this shadow memory to proactively assess the risk of pending actions prior to their execution. Extensive evaluation demonstrates that MAGE substantially outperforms existing defenses across diverse long-horizon threats in detection accuracy, achieves early-stage detection for the majority of attacks, and introduces only negligible overhead to agent utility. To our best knowledge, MAGE represents the first framework to detect and mitigate long-horizon threats using an agentic memory approach, establishing a new paradigm for this critical challenge and opening promising directions for future research.", "published": "2026-05-04", "updated": "2026-05-04", "authors": [ "Yuhui Wang", "Tanqiu Jiang", "Jiacheng Liang", "Charles Fleming", "Ting Wang" ], "categories": [ "cs.CR", "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2605.03228", "pdf_url": "https://arxiv.org/pdf/2605.03228", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.03312", "title": "MemFlow: Intent-Driven Memory Orchestration for Small Language Model Agents", "abstract": "Modern language agents must operate over long-horizon, multi-turn histories, yet deploying such agents with Small Language Models (SLMs) remains fundamentally difficult. Full-context prompting causes context overflow, flat retrieval exposes the model to noisy evidence, and open-ended agentic loops are unreliable under limited reasoning capacity. We argue that a substantial portion of SLM memory failure arises from mismatched memory operations: different query types demand categorically different retrieval strategies, evidence transformations, and context budgets that SLMs cannot reliably self-orchestrate through open-ended reasoning. We introduce MemFlow, a training-free memory orchestration framework that externalizes memory planning from the SLM. A Router Agent classifies each query by intent and dispatches it to the Memory Agent, which executes one of three specialized tiers (Profile Lookup, Targeted Retrieval, or Deep Reasoning) and assembles the resulting evidence under a dynamic, tier-aware token budget. An Answer Agent then generates a response from this compact context, and a Validator Agent optionally retries with a heavier memory tier when the response is not supported by the provided evidence. This route-then-compile design avoids tool-selection hallucination and reasoning loops while keeping the answer context compact. Evaluated on a frozen Qwen3-1.7B backbone across long-horizon memory benchmarks - LongMemEval, LoCoMo, and LongBench - MemFlow improves accuracy by nearly 2x over full-context SLM baselines. These results suggest that structured intent routing and deterministic evidence preparation can make limited-capacity models substantially more effective in resource-constrained long-horizon agents.", "published": "2026-05-05", "updated": "2026-05-05", "authors": [ "Jiayi Chen", "Yingcong Li", "Guiling Wang" ], "categories": [ "cs.MA" ], "url": "https://arxiv.org/abs/2605.03312", "pdf_url": "https://arxiv.org/pdf/2605.03312", "matched_queries": [ "agent-memory", "memory-evaluation" ] }, { "arxiv_id": "2605.03354", "title": "What Happens Inside Agent Memory? Circuit Analysis from Emergence to Diagnosis", "abstract": "Agent memory failures are silent: an LLM-based agent can produce a fluent response even when it fails to extract, retain, or retrieve the information needed across sessions. The write-manage-read loop describes the external pipeline of these systems but leaves open which internal computations implement each stage. Tracing feature circuits across the Qwen-3 family (0.6B--14B) and two memory frameworks (mem0 and A-MEM), we report two mechanistic findings and one deliverable. First, control is detectable before content: routing circuitry is causally active at 0.6B, while content circuitry produces no detectable signal until 4B, exposing a deployment regime where small models route memory decisions before they can reliably extract or ground the underlying facts. Second, the shared hub is recruited, not created: Write and Read converge on a late-layer hub that already exists in the base model as a context-grounding substrate, and memory framing recruits a memory-specific functional direction on this substrate rather than building one of its own. Both findings transfer across mem0 and A-MEM, indicating that the underlying computations are properties of the base model rather than of any particular interface. Building on this circuit structure, we develop an unsupervised stage-level diagnostic that localizes silent failures to the responsible operation up to 76.2% accuracy, outperforming the strongest supervised baseline by 13 points. Together, these results point to circuit-level signatures as a practical handle for monitoring and structurally-guided design of agent memory.", "published": "2026-05-05", "updated": "2026-05-07", "authors": [ "Xutao Mao", "Jinman Zhao", "Gerald Penn", "Cong Wang" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2605.03354", "pdf_url": "https://arxiv.org/pdf/2605.03354", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.03804", "title": "ScrapMem: A Bio-inspired Framework for On-device Personalized Agent Memory via Optical Forgetting", "abstract": "Long-term personalized memory for LLM agents is challenging on resource-limited edge devices due to high storage costs and multimodal complexity. To address this, we propose ScrapMem, a framework that integrates multimodal data into \"Scrapbook Page.\" ScrapMem introduces Optical Forgetting, an optical compression mechanism that progressively reduces the resolution of older memories, lowering storage cost while suppressing low-value details. To maintain semantic consistency, we construct an Episodic Memory Graph (EM-Graph) that organizes key events into a causal-temporal structure. Extensive experiments on the multimodal ATM-Bench showcase that ScrapMem provides three main benefits: (1) strong performance, achieving a new state-of-the-art with a 51.0% Joint@10 score; (2) high storage efficiency, reducing memory usage by up to 93% via optical forgetting; and (3) improved recall, increasing Recall@10 to 70.3% through structured aggregation. ScrapMem offers an effective and storage-efficient solution for on-device long-term memory in multimodal LLM agents.", "published": "2026-05-05", "updated": "2026-05-25", "authors": [ "Jiale Chang", "Yuxiang Ren" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2605.03804", "pdf_url": "https://arxiv.org/pdf/2605.03804", "matched_queries": [ "agent-memory", "episodic-memory", "long-term-memory" ] }, { "arxiv_id": "2605.06702", "title": "CASCADE: Case-Based Continual Adaptation for Large Language Models During Deployment", "abstract": "Large language models (LLMs) have become a central foundation of modern artificial intelligence, yet their lifecycle remains constrained by a rigid separation between training and deployment, after which learning effectively ceases. This limitation contrasts with natural intelligence, which continually adapts through interaction with its environment. In this paper, we formalise deployment-time learning (DTL) as the third stage in the LLM lifecycle that enables LLM agents to improve from experience during deployment without modifying model parameters. We present CASCADE (CASe-based Continual Adaptation during DEployment), a general and principled framework that equips LLM agents with an explicit, evolving episodic memory. CASCADE formulates experience reuse as a contextual bandit problem, enabling principled exploration-exploitation trade-offs and establishing no-regret guarantees over long-term interactions. This design allows agents to accumulate, select, and refine task-relevant cases, transforming past experience into actionable knowledge. Across 16 diverse tasks spanning medical diagnosis, legal analysis, code generation, web search, tool use, and embodied interaction, CASCADE improves macro-averaged success rate by 20.9% over zero-shot prompting while consistently outperforming gradient-based and memory-based baselines. By reframing deployment as an adaptive learning process, this work establishes a foundation for continually improving AI systems.", "published": "2026-05-05", "updated": "2026-05-05", "authors": [ "Siyuan Guo", "Yali Du", "Hechang Chen", "Yi Chang", "Jun Wang" ], "categories": [ "cs.AI", "cs.CL", "cs.LG" ], "url": "https://arxiv.org/abs/2605.06702", "pdf_url": "https://arxiv.org/pdf/2605.06702", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2605.04811", "title": "Tree-based Credit Assignment for Multi-Agent Memory System", "abstract": "Memory systems are widely adopted to enhance LLMs for long-horizon tasks, and are commonly organized as multi-agent pipelines with memory building, summarizing, and retrieval agents. To empower this system, existing RL-based methods either apply final downstream task rewards (e.g., QA accuracy) for all agents uniformly, which are coarse and ambiguous, or design task-specific rewards for agents on different subtasks, which require costly annotations (e.g., key evidence) and are difficult to define reliably. To address these limitations, we propose Tree-based Credit Assignment for Multi-Agent Memory Systems (TreeMem), which derives agent-specific credit from the final reward without task-specific annotations. Specifically, TreeMem extends the multi-agent pipeline (builder--summarizer--retrieval) into a tree structure, where each agent's outputs are expanded into multiple subsequent branches. The contribution of each agent is estimated via Monte Carlo averaging over its subsequent branches, capturing how intermediate agent actions may influence the final reward. This converts the coarse final reward into agent-specific optimization signals. These signals are then used to update all agent policies simultaneously, helping heterogeneous agents specialize effectively. Experiments on long-horizon benchmarks show that TreeMem improves memory system performance over strong baselines, validating the effectiveness of tree-structured credit assignment for the multi-agent memory system.", "published": "2026-05-06", "updated": "2026-05-06", "authors": [ "Marina Mao", "Alexandr Liu", "Pengbo Li", "Siheng Li", "Bo Zhou", "Xiang Wang" ], "categories": [ "cs.MA" ], "url": "https://arxiv.org/abs/2605.04811", "pdf_url": "https://arxiv.org/pdf/2605.04811", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.04897", "title": "Storage Is Not Memory: A Retrieval-Centered Architecture for Agent Recall", "abstract": "Extraction at ingestion is the wrong primitive for agent memory: content discarded before the query is known cannot be recovered at retrieval time. We propose True Memory, a six-layer architecture that shifts the center of the system from a storage schema to a multi-stage retrieval pipeline operating over events preserved verbatim. The full system runs as a single SQLite file on commodity CPU with no external database, vector index, graph store, or GPU. On LoCoMo (1,540 questions across 10 multi-session conversations), True Memory Pro reaches 93.0% accuracy (3-run mean) against 61.4% for Mem0, 65.4% for Supermemory, approximately 71% for Zep, and 94.5% for EverMemOS under a matched gpt-4.1-mini answer model. On LongMemEval (500 questions), True Memory Pro reaches 87.8% (3-run mean). On BEAM-1M (700 questions at the 1-million-token scale), True Memory Pro reaches 76.6% (3-run mean), above the prior published result of 73.9% for Hindsight. A 56-configuration ablation shows a 1.3-percentage-point spread within the top-performing configuration family.", "published": "2026-05-06", "updated": "2026-05-06", "authors": [ "Joshua Adler", "Guy Zehavi" ], "categories": [ "cs.CL", "cs.AI", "cs.IR" ], "url": "https://arxiv.org/abs/2605.04897", "pdf_url": "https://arxiv.org/pdf/2605.04897", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.05583", "title": "Belief Memory: Agent Memory Under Partial Observability", "abstract": "LLM agents that operate over long context depend on external memory to accumulate knowledge over time. However, existing methods typically store each observation as a single deterministic conclusion (e.g., inferring \"API~X failed\" from temporary errors), even though such observations are inherently partial and potentially ambiguous. By committing to one conclusion and discarding uncertainty, these methods introduce self-reinforcing error: the agent acts on the stored conclusion, never revisits alternatives, and reinforces the conclusion over time. To address this issue, we propose BeliefMem, which shifts the memory paradigm from committing to a single conclusion per observation to retaining multiple candidate conclusions with their probabilities. Concretely, BeliefMem stores the candidate conclusions as separate memory entries, each carrying a probability that is updated via Noisy-OR rules as new observations arrive. At retrieval, all candidates surface together with their probabilities, keeping alternatives visible to the agent. Since each conclusion in memory retains its probability, BeliefMem preserves the uncertainty that the deterministic paradigm discards, enabling the agent to act with high confidence on well-evidenced knowledge while retaining the capacity to update its confidence when new evidence arrives. Empirical evaluations on LoCoMo and ALFWorld benchmarks show that, even with limited data, BeliefMem achieves the best average performance, remarkably outperforming well-known baselines. More broadly, such probabilistic memory produces substantial gains and explores a new direction for agent memory in partially observable environments.", "published": "2026-05-07", "updated": "2026-05-08", "authors": [ "Junfeng Liao", "Qizhou Wang", "Jianing Zhu", "Bo Du", "Rui Yan", "Xiuying Chen" ], "categories": [ "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2605.05583", "pdf_url": "https://arxiv.org/pdf/2605.05583", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.06132", "title": "MemReranker: Reasoning-Aware Reranking for Agent Memory Retrieval", "abstract": "In agent memory systems, the reranking model serves as the critical bridge connecting user queries with long-term memory. Most systems adopt the \"retrieve-then-rerank\" two-stage paradigm, but generic reranking models rely on semantic similarity matching and lack genuine reasoning capabilities, leading to a problem where recalled results are semantically highly relevant yet do not contain the key information needed to answer the question. This deficiency manifests in memory scenarios as three specific problems. First, relevance scores are miscalibrated, making threshold-based filtering difficult. Second, ranking degrades when facing temporal constraints, causal reasoning, and other complex queries. Third, the model cannot leverage dialogue context for semantic disambiguation. This report introduces MemReranker, a reranking model family (0.6B/4B) built on Qwen3-Reranker through multi-stage LLM knowledge distillation. Multi-teacher pairwise comparisons generate calibrated soft labels, BCE pointwise distillation establishes well-distributed scores, and InfoNCE contrastive learning enhances hard-sample discrimination. Training data combines general corpora with memory-specific multi-turn dialogue data covering temporal constraints, causal reasoning, and coreference resolution. On the memory retrieval benchmark, MemReranker-0.6B substantially outperforms BGE-Reranker and matches open-source 4B/8B models as well as GPT-4o-mini on key metrics. MemReranker-4B further achieves 0.737 MAP, with several metrics on par with Gemini-3-Flash, while maintaining inference latency at only 10--20% of large models. On finance and healthcare vertical-domain benchmarks, the models preserve generalization capabilities on par with mainstream large-parameter rerankers.", "published": "2026-05-07", "updated": "2026-05-14", "authors": [ "Chunyu Li", "Mengyuan Zhang", "Jingyi Kang", "Ding Chen", "Jiajun Shen", "Bo Tang", "Xuanhe Zhou", "Feiyu Xiong", "Zhiyu Li" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2605.06132", "pdf_url": "https://arxiv.org/pdf/2605.06132", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.06527", "title": "STALE: Can LLM Agents Know When Their Memories Are No Longer Valid?", "abstract": "Large Language Model (LLM) agents are increasingly expected to maintain coherent, long-term personalized memory, yet current benchmarks primarily measure static fact retrieval, overlooking the ability to revise stored beliefs when new evidence emerges. We identify a critical and underexplored failure mode, Implicit Conflict: a later observation invalidates an earlier memory without explicit negation, requiring contextual inference and commonsense reasoning to detect. To rigorously evaluate this capability, we introduce STALE, a benchmark of 400 expert-validated conflict scenarios (1,200 evaluation queries across three probing dimensions) spanning over 100 everyday topics with contexts up to 150K tokens. We propose a three-dimensional probing framework that tests State Resolution (detecting that a prior belief is outdated), Premise Resistance (rejecting queries that falsely presuppose a stale state), and Implicit Policy Adaptation (proactively applying updated states in downstream behavior). A systematic evaluation of frontier LLMs and specialized memory frameworks reveals a pervasive gap between retrieving updated evidence and acting on it, with even the best evaluated model achieving only 55.2% overall accuracy. Models often accept outdated assumptions embedded in a user's query, and they struggle to recognize when a change in one aspect of the user's state should invalidate related memories. To establish an initial baseline for state-aware memory, we further present CUPMem, a prototype that strengthens write-time revision through structured state consolidation and propagation-aware search, suggesting that explicit state adjudication is a promising direction for robust agentic memory.", "published": "2026-05-07", "updated": "2026-05-07", "authors": [ "Hanxiang Chao", "Yihan Bai", "Rui Sheng", "Tianle Li", "Yushi Sun" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2605.06527", "pdf_url": "https://arxiv.org/pdf/2605.06527", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.06716", "title": "From Storage to Experience: A Survey on the Evolution of LLM Agent Memory Mechanisms", "abstract": "Large Language Model (LLM)-based agents have fundamentally reshaped artificial intelligence by integrating external tools and planning capabilities. While memory mechanisms have emerged as the architectural cornerstone of these systems, current research remains fragmented, oscillating between operating system engineering and cognitive science. This theoretical divide prevents a unified view of technological synthesis and a coherent evolutionary perspective. To bridge this gap, this survey proposes a novel evolutionary framework for LLM agent memory mechanisms, formalizing the development process into three stages: Storage (trajectory preservation), Reflection (trajectory refinement), and Experience (trajectory abstraction). We first formally define these three stages before analyzing the three core drivers of this evolution: the necessity for long-range consistency, the challenges in dynamic environments, and the ultimate goal of continual learning. Furthermore, we specifically explore two transformative mechanisms in the frontier Experience stage: proactive exploration and cross-trajectory abstraction. By synthesizing these disparate views, this work offers robust design principles and a clear roadmap for the development of next-generation LLM agents.", "published": "2026-05-07", "updated": "2026-05-07", "authors": [ "Jinghao Luo", "Yuchen Tian", "Chuxue Cao", "Ziyang Luo", "Hongzhan Lin", "Kaixin Li", "Chuyi Kong", "Ruichao Yang", "Jing Ma" ], "categories": [ "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2605.06716", "pdf_url": "https://arxiv.org/pdf/2605.06716", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.06812", "title": "Towards Security-Auditable LLM Agents: A Unified Graph Representation", "abstract": "LLM-based agentic systems are rapidly evolving to perform complex autonomous tasks through dynamic tool invocation, stateful memory management, and multi-agent collaboration. However, this semantics-driven execution paradigm creates a severe semantic gap between low-level physical events and high-level execution intent, making post-hoc security auditing fundamentally difficult. Existing representation mechanisms, including static SBOMs and runtime logs, provide only fragmented evidence and fail to capture cognitive-state evolution, capability bindings, persistent memory contamination, and cascading risk propagation across interacting agents. To bridge this gap, we propose Agent-BOM, a unified structural representation for agent security auditing. Agent-BOM models an agentic system as a hierarchical attributed directed graph that separates static capability bases, such as models, tools, and long-term memory, from dynamic runtime semantic states, such as goals, reasoning trajectories, and actions. These layers are connected through semantic edges and security attributes, transforming fragmented execution traces into queryable audit paths. Building on Agent-BOM, we develop a graph-query-based paradigm for path-level risk assessment and instantiate it with the OWASP Agentic Top 10. We further implement an auditing plugin in the OpenClaw environment to construct Agent-BOM from live executions. Evaluation on representative real-world agentic attack scenarios shows that Agent-BOM can reconstruct stealthy attack chains, including cross-session memory poisoning and tool misuse, capability supply-chain hijacking and unexpected code execution, multi-agent ecosystem hijacking, and privilege and trust abuse. These results demonstrate that Agent-BOM provides a unified and auditable foundation for root-cause analysis and security adjudication in complex agentic ecosystems.", "published": "2026-05-07", "updated": "2026-05-07", "authors": [ "Chaofan Li", "Lyuye Zhang", "Jintao Zhai", "Siyue Feng", "Xichun Yang", "Huahao Wang", "Shihan Dou", "Yu Ji", "Yutao Hu", "Yueming Wu", "Yang Liu", "Deqing Zou" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2605.06812", "pdf_url": "https://arxiv.org/pdf/2605.06812", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2605.07242", "title": "MEMOREPAIR: Barrier-First Cascade Repair in Agentic Memory", "abstract": "Agentic memory evolves across tasks into durable derived artifacts: summaries, cached outputs, embeddings, learned skills, and executable tool procedures. When a source artifact is deleted, corrected, or invalidated by tool or API migration, descendants derived from that source can remain visible and steer future actions with stale support. We formalize this failure mode as the cascade update problem, where repair targets the visible derived state of the memory store. We present MemoRepair, a barrier-first cascade-repair contract for agentic memory. A repair event induces a controlled transition from invalidated descendant state to validated successor state: affected descendants are withdrawn before repair, successors are constructed from retained support and staged repaired predecessors under the current interface, and republication is restricted to validated predecessor-closed successors. This contract induces a scalarized repair-selection problem for a fixed repair-cost tradeoff. We show that the induced publication problem reduces to maximum-weight predecessor closure and can be solved exactly by a single s-t min-cut. Experiments on ToolBench and MemoryArena show that, with complete influence provenance, MemoRepair reduces invalidated-memory exposure from 69.8-94.3% under systems without cascade repair to 0%. Compared with exhaustive Repair all, it recovers 91.1-94.3% of validated successors while reducing normalized repair-operator cost from 1.00 to 0.57-0.76.", "published": "2026-05-08", "updated": "2026-05-08", "authors": [ "Yang Zhao", "Chengxiao Dai", "Mengying Kou", "Yue Xiu" ], "categories": [ "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2605.07242", "pdf_url": "https://arxiv.org/pdf/2605.07242", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.07313", "title": "When Stored Evidence Stops Being Usable: Scale-Conditioned Evaluation of Agent Memory", "abstract": "Memory-agent evaluations report fixed-snapshot accuracy or retrieval quality, but these scores do not show whether evidence remains usable as irrelevant sessions (sessions not annotated as task-relevant evidence for the query) accumulate. We present a scale-conditioned evaluation protocol for agent memory under evidence-preserving growth: for each query, task evidence is held fixed while irrelevant sessions are added. The protocol logs agent--memory trajectories and reports four diagnostics: budget-compliant reliability, tail memory-call burden, failure-regime decomposition, and the usable-scale boundary where reliability falls below the target. Applied to LongMemEval and LoCoMo across flat, planar, and hierarchical memory interfaces, the protocol shows reliability loss is not a single phenomenon. On LongMemEval, HippoRAG stays within the two-call budget but loses 16--20 percentage points in budget-compliant reliability as irrelevant sessions are added; LiCoMemory's observed failures depend strongly on the agent, with Qwen3-8B exceeding the budget while Qwen3-32B and Qwen3-235B remain reliable in the tested range. The result supports a framework for making scalable-memory claims conditional on agent, interface, scale range, and interaction budget.", "published": "2026-05-08", "updated": "2026-05-08", "authors": [ "Jiaqi Shao", "Yiyi Lu", "Yunzhen Zhang", "Bing Luo" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2605.07313", "pdf_url": "https://arxiv.org/pdf/2605.07313", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.08374", "title": "MemQ: Integrating Q-Learning into Self-Evolving Memory Agents over Provenance DAGs", "abstract": "Episodic memory allows LLM agents to accumulate and retrieve experience, but current methods treat each memory independently, i.e., evaluating retrieval quality in isolation without accounting for the dependency chains through which memories enable the creation of future memories. We introduce MemQ, which applies TD($λ$) eligibility traces to memory Q-values, propagating credit backward through a provenance DAG that records which memories were retrieved when each new memory was created. Credit weight decays as $(γλ)^d$ with DAG depth $d$, replacing temporal distance with structural proximity. We formalize the setting as an Exogenous-Context MDP, whose factored transition decouples the exogenous task stream from the endogenous memory store. Across six benchmarks, spanning OS interaction, function calling, code generation, multimodal reasoning, embodied reasoning, and expert-level QA, MemQ achieves the highest success rate on all six in generalization evaluation and runtime learning, with gains largest on multi-step tasks that produce deep and relevant provenance chains (up to +5.7~pp) and smallest on single-step classification (+0.77~pp) where single-step updates already suffice. We further study how $γ$ and $λ$ interact with the EC-MDP structure, providing principled guidance for parameter selection and future research. Code is available at https://github.com/jwliao-ai/MemQ.", "published": "2026-05-08", "updated": "2026-05-14", "authors": [ "Junwei Liao", "Haoting Shi", "Ruiwen Zhou", "Jiaqian Wang", "Shengtao Zhang", "Wei Zhang", "Ying Wen", "Zhiyu Li", "Feiyu Xiong", "Bo Tang", "Weinan Zhang", "Muning Wen" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2605.08374", "pdf_url": "https://arxiv.org/pdf/2605.08374", "matched_queries": [ "agent-memory", "episodic-memory" ] }, { "arxiv_id": "2605.08442", "title": "Defense effectiveness across architectural layers: a mechanistic evaluation of persistent memory attacks on stateful LLM agents", "abstract": "Persistent memory in LLM agents creates an attack surface that production safety classifiers do not observe: the payload enters via RAG retrieval and persists across sessions via tool-mediated memory. We evaluate six defenses across four architectural layers against delayed-trigger attacks on nine open-source models (5,040 runs, N=40 per condition). Five of six defenses fail: input-level filters never see the payload (it enters via RAG, not user input); retrieval-level classifiers observe it but cannot distinguish compliance-framed injection from legitimate policy; instruction-level hardening is overridden by the stored rule's compliance framing. Only tool-gating at the memory layer (Memory Sandbox) reduces ASR to 0% for eight of nine models, with zero utility cost. A reasoning model inverts this defense via goal-directed RAG fallback, a mechanism that replicates cross-family on Bedrock. A reasoning-mode ablation reveals a double dissociation: no single sandbox implementation is safe across both reasoning and non-reasoning model classes. We resolve this with a content-layer proof-of-concept (RATG), validated on non-reasoning models. A loaded-corpus frontier evaluation (21 models, 3 providers, N=40) overturns an initial empty-corpus screen showing 0/210 exfiltrations: that was a threat-model artifact, not model safety. Under realistic conditions, Gemini 3.1 Pro Preview exfiltrates at 95% ASR, GPT-5.1 regresses to 22.5% relative to GPT-5 (5%), and Anthropic blocks at the injection layer (0-17.5% storage, 0% ASR). Nearly all OpenAI and Gemini models store the rule at 100% regardless of execution resistance, creating supply-chain risk in shared-memory deployments. Defense effectiveness is determined by architectural layer and reasoning capability, not classifier quality.", "published": "2026-05-08", "updated": "2026-07-03", "authors": [ "Jun Wen Leong" ], "categories": [ "cs.CR", "cs.AI", "cs.LG" ], "url": "https://arxiv.org/abs/2605.08442", "pdf_url": "https://arxiv.org/pdf/2605.08442", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2605.08468", "title": "PYTHALAB-MERA: Validation-Grounded Memory, Retrieval, and Acceptance Control for Frozen-LLM Coding Agents", "abstract": "Local LLM-based coding agents increasingly work in settings where correctness is earned through execution feedback, persistent state, and bounded repair, not through a single fluent answer. Static retrieval, long-context prompting, self-refinement, execution-feedback repair, and reinforcement learning over model weights each address part of this setting, but they do not jointly provide validation-grounded episodic memory, adaptive retrieval-action selection, delayed credit assignment, and structural skill reuse around a frozen local model. We introduce PYTHALAB-MERA, a lightweight external controller for local validation-conditioned code generation. The frozen language model proposes complete source files; the controller decides which memory records and AST-derived skills should enter the next prompt, validates each candidate through a fail-fast pipeline, converts validation outcomes into bounded shaped rewards, and propagates delayed credit through TD(lambda)-style eligibility traces. We evaluate the implementation as a local CLI artifact on reinforcement-learning coding tasks with strict validation gates. In the measured hard RL setting with three tasks, three repetitions, and a three-attempt budget, PYTHALAB-MERA passed 8/9 strict validations; the self-refinement baseline and the investigated GRACE extension each passed 0/9. These results support a deliberately bounded claim: in this recorded setting, the external memory-and-retrieval controller improved validation success. They do not establish general-purpose code synthesis, state-of-the-art performance, formal program correctness, or formal safety.", "published": "2026-05-08", "updated": "2026-05-08", "authors": [ "Mehmet Iscan" ], "categories": [ "cs.CL", "cs.AI", "cs.LG" ], "url": "https://arxiv.org/abs/2605.08468", "pdf_url": "https://arxiv.org/pdf/2605.08468", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2605.08538", "title": "Human-Inspired Memory Architecture for LLM Agents", "abstract": "Current LLM agents lack principled mechanisms for managing persistent memory across long interaction horizons. We present a biologically-grounded memory architecture comprising six cognitive mechanisms: (1) sleep-phase consolidation, (2) interference-based forgetting, (3) engram maturation, (4) reconsolidation upon retrieval, (5) entity knowledge graphs, and (6) hybrid multi-cue retrieval. Each mechanism addresses a specific failure mode of naive memory accumulation. We introduce a synthetic calibration methodology that derives all pipeline thresholds without benchmark data exposure, eliminating a common source of evaluation leakage. We evaluate on two benchmarks. First, a VSCode issue-tracking dataset (13K issues, 120K events) where deduplication-based consolidation achieves 97.2% retention precision with 58% store reduction (+21.8 pp over baseline). Second, the LongMemEval personal-chat benchmark where we conduct the first streaming M-tier evaluation (475 sessions, ~540K unique turns). At a 200K-token context budget, our pipeline matches raw retrieval accuracy (70.1% vs. 71.2%, overlapping 95% CI) while exposing a tunable accuracy/store-size operating curve. At S-tier scale (50 sessions), dedup-based consolidation yields a +13.3 pp improvement in preference recall.", "published": "2026-05-08", "updated": "2026-05-08", "authors": [ "Doga Kerestecioglu", "Alexei Robsky", "Clemens Vasters", "Anshul Sharma", "Yitzhak Kesselman" ], "categories": [ "cs.AI", "cs.CL", "cs.IR", "cs.LG" ], "url": "https://arxiv.org/abs/2605.08538", "pdf_url": "https://arxiv.org/pdf/2605.08538", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2605.09033", "title": "ShadowMerge: A Novel Poisoning Attack on Graph-Based Agent Memory via Relation-Channel Conflicts", "abstract": "Graph-based agent memory is increasingly used in LLM agents to support structured long-term recall and multi-hop reasoning, but it also creates a new poisoning surface: an attacker can inject a crafted relation into graph memory so that it is later retrieved and influences agent behavior. Existing agent-memory poisoning attacks mainly target flat textual records and are ineffective in graph-based memory because malicious relations often fail to be extracted, merged into the target anchor neighborhood, or retrieved for the victim query. We present SHADOWMERGE, a poisoning attack against graph-based agent memory that exploits relation-channel conflicts. Its key insight is that a poisoned relation can share the same query-activated anchor and canonicalized relation channel as benign evidence while carrying a conflicting value. To realize this, we design AIR, a pipeline that converts the conflict into an ordinary interaction that can be extracted, merged, and retrieved by the graph-memory system. We evaluate SHADOWMERGE on Mem0 and three public real-world datasets: PubMedQA, WebShop, and ToolEmu. SHADOWMERGE achieves 93.8% average attack success rate, improving the best baseline by 50.3 absolute points, while having negligible impact on unrelated benign tasks. Mechanism studies show that SHADOWMERGE overcomes the three key limitations of existing agent-memory poisoning attacks, and defense analysis shows that representative input-side defenses are insufficient to mitigate it. We have responsibly disclosed our findings to affected graph-memory vendors and open sourced SHADOWMERGE.", "published": "2026-05-09", "updated": "2026-05-15", "authors": [ "Yang Luo", "Zifeng Kang", "Tiantian Ji", "Xinran Liu", "Yong Liu", "Shuyu Li", "Lingyun Peng" ], "categories": [ "cs.CR", "cs.AI" ], "url": "https://arxiv.org/abs/2605.09033", "pdf_url": "https://arxiv.org/pdf/2605.09033", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.09315", "title": "Do Self-Evolving Agents Forget? Capability Degradation and Preservation in Lifelong LLM Agent Adaptation", "abstract": "Recent advances in LLM agents enable systems that autonomously refine workflows, accumulate reusable skills, self-train their underlying models, and maintain persistent memory. However, we show that such self-evolution is often non-monotonic: adapting to new task distributions can progressively degrade previously acquired capabilities across all major evolution channels. We identify this phenomenon as \\emph{capability erosion under self-evolution} and show that it consistently emerges across workflow, skill, model, and memory evolution. To mitigate this issue, we propose \\emph{Capability-Preserving Evolution} (CPE), a general stabilization principle that constrains destructive capability drift during continual adaptation. Across all four evolution dimensions, CPE consistently improves retained capability stability while preserving adaptation performance. For example, in workflow evolution, CPE improves retained simple-task performance from 41.8\\% to 52.8\\% under GPT-5.1 optimization while simultaneously achieving stronger complex-task adaptation. Our findings suggest that stable long-horizon self-evolving agents require not only acquiring new capabilities, but also explicitly preserving previously learned ones during continual adaptation.", "published": "2026-05-10", "updated": "2026-05-10", "authors": [ "Ye Yu", "Xiaopeng Yuan", "Haibo Jin", "Heming Liu", "Yaoning Yu", "Haohan Wang" ], "categories": [ "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2605.09315", "pdf_url": "https://arxiv.org/pdf/2605.09315", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2605.09330", "title": "The Trap of Trajectory: Towards Understanding and Mitigating Spurious Correlations in Agentic Memory", "abstract": "Agentic memory enables LLMs to persist information beyond a single context window and reuse it in later decisions, but it also introduces a new vulnerability: spurious correlations, where retrieved memory carries miscorrelated evidence and propagates erroneous reasoning into downstream decisions. Despite the widespread use of agentic memory, this risk remains largely underexplored. We address it from two aspects. First, we benchmark several canonical types of spurious patterns identified through causal structure and record them across trajectory-level memory. Diagnosing agentic memory systems on this benchmark reveals that memory improves reasoning on clean inputs but amplifies reliance on spurious patterns when they are present. Second, we propose CAMEL, a plug-and-play calibration method that operates across diverse memory architectures at both write and retrieval time. CAMEL consistently reduces reliance on spurious patterns across all three types while preserving or improving performance on clean inputs and staying robust under adaptive attacks targeting the calibration. Overall, CAMEL offers a principled and lightweight solution toward more reliable agentic memory deployment.", "published": "2026-05-10", "updated": "2026-05-10", "authors": [ "Luoxi Tang", "Rupali Rajendra Vaje", "Yuqiao Meng", "Sakshi Sunil Narkar", "Weicheng Ma", "Zeyu Ding", "Dazheng Zhang", "Zhaohan Xi" ], "categories": [ "cs.LG", "cs.AI" ], "url": "https://arxiv.org/abs/2605.09330", "pdf_url": "https://arxiv.org/pdf/2605.09330", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.11032", "title": "Portable Agent Memory: A Protocol for Cryptographically-Verified Memory Transfer Across Heterogeneous AI Agents", "abstract": "We present Portable Agent Memory, an open protocol and reference implementation for transferring persistent memory state across heterogeneous AI agents. Modern AI agents accumulate rich context -- episodic events,semantic knowledge, procedural skills, working state, and identity preferences -- but this context remains locked within vendor-specific runtimes. Portable Agent Memory addresses this through: (1) a five-component structured memory model with content-addressable entries linked by a Merkle-DAG provenance graph providing tamper-evidence; (2) capability-based access control enabling selective, scoped disclosure of memory segments; (3) an injection-resistant rehydration protocol that adapts recalled content to heterogeneous target models while mitigating indirect prompt injection; and (4) a JSON-first serialization format with optional CBOR compaction for efficient transport. We provide a Python SDK with 54 passing tests, agent skills for multiple platforms, and demonstrate cross-model memory transfer between GPT-4, Claude, Gemini, and Llama architectures. The protocol is open-source under Apache 2.0.", "published": "2026-05-10", "updated": "2026-05-10", "authors": [ "Santhosh Kumar Ravindran" ], "categories": [ "cs.CR", "cs.AI" ], "url": "https://arxiv.org/abs/2605.11032", "pdf_url": "https://arxiv.org/pdf/2605.11032", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.09863", "title": "Nautilus Compass: Black-box Persona Drift Detection for Production LLM Agents", "abstract": "Production LLM coding agents drift over long sessions: they forget user-specified constraints, slip into mistakes the user already flagged, and confabulate prior agreements. White-box approaches such as persona vectors require model weights and so cannot be applied to closed APIs (Claude, GPT-4) that most users actually interact with. We present Nautilus Compass, a black-box persona drift detector and agent memory layer for production coding agents. The method operates entirely at the prompt-text layer: cosine similarity between user prompts and behavioral anchor texts, aggregated by a weighted top-k mean using BGE-m3 embeddings. Compass is, to our knowledge, the only public agent memory layer (among Mem0, Letta, Cognee, Zep, MemOS, smrti verified May 2026) that does not call an LLM at index time to extract facts or build a graph; raw conversation text is embedded directly. The system ships as a Claude Code plugin, an MCP 2024-11-05 A2A server (Cursor, Cline, Hermes), a CLI, and a REST API on one daemon, with a Merkle-chained audit log for tamper-evident anchor updates. On a held-out test set built from real Claude Code session traces and labeled by an independent LLM judge, Compass reaches ROC AUC 0.83 for drift detection. The embedded retrieval pipeline scores 56.6% on LongMemEval-S v0.8 and 44.4% on EverMemBench-Dynamic (n=500), topping the four published EverMemBench Table 4 baselines. LongMemEval-S 56.6% is ~30 points below recent white-box leaders (90+%); we treat that as the architectural ceiling of the no-extraction design. End-to-end reproduction cost is $3.50 (~14x cheaper than GPT-4o-judged stacks). A paired cross-vendor behavior A/B accompanies these numbers as preliminary system-level evidence. Code, anchors, frozen test data, and audit-log tooling are MIT-licensed at github.com/chunxiaoxx/nautilus-compass.", "published": "2026-05-11", "updated": "2026-05-11", "authors": [ "Chunxiao Wang" ], "categories": [ "cs.CR", "cs.AI", "cs.CL", "cs.IR", "cs.LG" ], "url": "https://arxiv.org/abs/2605.09863", "pdf_url": "https://arxiv.org/pdf/2605.09863", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.09942", "title": "HAGE: Harnessing Agentic Memory via RL-Driven Weighted Graph Evolution", "abstract": "Memory retrieval in agentic large language model (LLM) systems is often treated as a static lookup problem, relying on flat vector search or fixed binary relational graphs. However, fixed graph structures cannot capture the varying strength, confidence, and query-dependent relevance of relationships between events. In this paper, we propose HAGE, a weighted multi-relational memory framework that reconceptualizes retrieval as sequential, query-conditioned traversal over a unified relational memory graph. Memory is organized as relation-specific graph views over shared memory nodes, where each edge is associated with a trainable relation feature vector encoding multiple relational signals. Given a query, an LLM-based classifier identifies the relational intent, and a routing network dynamically modulates the corresponding dimensions of the edge embedding. Traversal scores are computed via a learned combination of semantic similarity and these query-conditioned edge representations. This allows memory traversal to prioritize high-utility relational paths while softly suppressing noisy or weakly relevant connections. Beyond adaptive traversal, HAGE further introduces a reinforcement learning-based training framework that jointly optimizes routing behavior and edge representations using downstream tasks. Finally, empirical results demonstrate improved long-horizon reasoning accuracy and a favorable accuracy-efficiency trade-off compared to state-of-the-art agentic memory systems. Our code is available at https://github.com/FredJiang0324/HAGE_MVPReview.", "published": "2026-05-11", "updated": "2026-05-11", "authors": [ "Dongming Jiang", "Yi Li", "Guanpeng Li", "Qiannan Li", "Bingzhe Li" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2605.09942", "pdf_url": "https://arxiv.org/pdf/2605.09942", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.10268", "title": "MemReread: Enhancing Agentic Long-Context Reasoning via Memory-Guided Rereading", "abstract": "To tackle long-context reasoning tasks without the quadratic complexity of standard attention mechanisms, approaches based on agent memory have emerged, which typically maintain a dynamically updated memory when linearly processing document chunks. To mitigate the potential loss of latent evidence in this memorize-while-reading paradigm, recent works have integrated retrieval modules that allow agents to recall information previously discarded during memory overwriting. However, retrieval-based recall suffers from both evidence loss during memory formation and interference induced by invalid queries. To overcome these limitations, we propose MemReread. Built upon streaming reading, MemReread circumvents intermediate retrieval. It triggers question decomposition and rereading when the final memory is insufficient, enabling the recovery of indirect facts that were prematurely discarded. This design supports non-linear reasoning while preserving the inherent logical flow of document comprehension. To further enhance practicality, we introduce a reinforcement learning framework that enhances length extrapolation capability while dynamically determining the number of rereading passes based on task complexity, thereby flexibly controlling computational overhead. Extensive experiments demonstrate that MemReread consistently outperforms baseline frameworks on long-context reasoning tasks, while maintaining linear time complexity with respect to context length.", "published": "2026-05-11", "updated": "2026-05-11", "authors": [ "Baibei Ji", "Xiaoyang Weng", "Juntao Li", "Zecheng Tang", "Yihang Lou", "Min Zhang" ], "categories": [ "cs.CL", "cs.AI" ], "url": "https://arxiv.org/abs/2605.10268", "pdf_url": "https://arxiv.org/pdf/2605.10268", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.10870", "title": "Remember the Decision, Not the Description: A Rate-Distortion Framework for Agent Memory", "abstract": "Long-horizon language agents must operate under limited runtime memory, yet existing memory mechanisms often organize experience around descriptive criteria such as relevance, salience, or summary quality. For an agent, however, memory is valuable not because it faithfully describes the past, but because it preserves the distinctions between histories that must remain separated under a fixed budget to support good decisions. We cast this as a decision-centric rate-distortion problem, measuring memory quality by the loss in achievable decision quality induced by compression. This yields an exact forgetting boundary for what can be safely forgotten, and a memory-distortion frontier characterizing the optimal tradeoff between memory budget and decision quality. Motivated by this decision-centric view of memory, we propose DeMem, an online memory learner that refines its partition only when data certify that a shared state would induce decision conflict, and prove near-minimax regret guarantees. On both controlled synthetic diagnostics and long-horizon conversational benchmarks, DeMem yields consistent gains under the same runtime budget, supporting the principle that memory should preserve the distinctions that matter for decisions, not descriptions.", "published": "2026-05-11", "updated": "2026-05-11", "authors": [ "Mingxi Zou", "Zhihan Guo", "Langzhang Liang", "Zhuo Wang", "Qifan Wang", "Qingsong Wen", "Irwin King", "Lizhen Qu", "Zenglin Xu" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2605.10870", "pdf_url": "https://arxiv.org/pdf/2605.10870", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.10899", "title": "RubricEM: Meta-RL with Rubric-guided Policy Decomposition beyond Verifiable Rewards", "abstract": "Training deep research agents, namely systems that plan, search, evaluate evidence, and synthesize long-form reports, pushes reinforcement learning beyond the regime of verifiable rewards. Their outputs lack ground-truth answers, their trajectories span many tool-augmented decisions, and standard post-training offers little mechanism for turning past attempts into reusable experience. In this work, we argue that rubrics should serve not merely as final-answer evaluators, but as the shared interface that structures policy execution, judge feedback, and agent memory. Based on this view, we introduce RubricEM, a rubric-guided reinforcement learning framework that combines stagewise policy decomposition with reflection-based meta-policy evolution. RubricEM first makes research trajectories stage-aware by conditioning planning, evidence gathering, review, and synthesis on self-generated rubrics. It then assigns credit with Stage-Structured GRPO, which uses stagewise rubric judgments to provide denser semantic feedback for long-horizon optimization. In parallel, RubricEM trains a shared-backbone reflection meta-policy that distills judged trajectories into reusable rubric-grounded guidance for future attempts. The resulting RubricEM-8B achieves strong performance across four long-form research benchmarks, outperforming comparable open models and approaching proprietary deep-research systems. Beyond final performance, we perform thorough analyses to understand the key ingredients of RubricEM.", "published": "2026-05-11", "updated": "2026-05-11", "authors": [ "Gaotang Li", "Bhavana Dalvi Mishra", "Zifeng Wang", "Jun Yan", "Yanfei Chen", "Chun-Liang Li", "Long T. Le", "Rujun Han", "George Lee", "Hanghang Tong", "Chen-Yu Lee", "Tomas Pfister" ], "categories": [ "cs.CL", "cs.LG" ], "url": "https://arxiv.org/abs/2605.10899", "pdf_url": "https://arxiv.org/pdf/2605.10899", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.13880", "title": "PREPING: Building Agent Memory without Tasks", "abstract": "Agent memory is typically constructed either offline from curated demonstrations or online from post-deployment interactions. However, regardless of how it is built, an agent faces a cold-start gap when first introduced to a new environment without any task-specific experience available. In this paper, we study pre-task memory construction: whether an agent can build procedural memory before observing any target-environment tasks, using only self-generated synthetic practice. Yet, synthetic interaction alone is insufficient, as without controlling what to practice and what to store, synthetic tasks become redundant, infeasible, and ultimately uninformative, and memory further degrades quickly due to unfiltered trajectories. To overcome this, we present Preping, a proposer-guided memory construction framework. At its core is proposer memory, a structured control state that shapes future practice. A Proposer generates synthetic tasks conditioned on this state, a Solver executes them, and a Validator determines which trajectories are eligible for memory insertion while also providing feedback to guide future proposals. Experiments on AppWorld, BFCL v3, and MCP-Universe show that Preping substantially improves over a no-memory baseline and achieves performance competitive with strong playbook-based methods built from offline or online experience, with deployment cost $2.99\\times$ lower on AppWorld and $2.23\\times$ lower on BFCL v3 than online memory construction. Further analyses reveal that the main benefit does not come from synthetic volume alone, but from proposer-side control over feasibility, redundancy, and coverage, combined with selective memory updates.", "published": "2026-05-11", "updated": "2026-05-11", "authors": [ "Yumin Choi", "Sangwoo Park", "Minki Kang", "Jinheon Baek", "Sung Ju Hwang" ], "categories": [ "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2605.13880", "pdf_url": "https://arxiv.org/pdf/2605.13880", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.11814", "title": "MedMemoryBench: Benchmarking Agent Memory in Personalized Healthcare", "abstract": "The large-scale deployment of personalized healthcare agents demands memory mechanisms that are exceptionally precise, safe, and capable of long-term clinical tracking. However, existing benchmarks primarily focus on daily open-domain conversations, failing to capture the high-stakes complexity of real-world medical applications. Motivated by the stringent production requirements of an industry-leading health management agent serving tens of millions of active users, we introduce MedMemoryBench. We develop a human-agent collaborative pipeline to synthesize highly realistic, long-horizon medical trajectories based on clinically grounded, synthetic patient archetypes. This process yields a massive, expertly validated dataset comprising approximately 2,000 sessions and 16,000 interaction turns. Crucially, MedMemoryBench departs from traditional static evaluations by pioneering an \"evaluate-while-constructing\" streaming assessment protocol, which precisely mirrors dynamic memory accumulation in production environments. Furthermore, we formalize and systematically investigate the critical phenomenon of memory saturation, where sustained information influx actively degrades retrieval and reasoning robustness. Comprehensive benchmarking reveals severe bottlenecks in mainstream architectures, particularly concerning complex medical reasoning and noise resilience. By exposing these fundamental flaws, MedMemoryBench establishes a vital foundation for developing robust, production-ready medical agents.", "published": "2026-05-12", "updated": "2026-05-12", "authors": [ "Yihao Wang", "Haoran Xu", "Renjie Gu", "Yixuan Ye", "Xinyi Chen", "Xinyu Mu", "Yuan Gao", "Chunxiao Guo", "Peng Wei", "Jinjie Gu", "Huan Li", "Ke Chen", "Lidan Shou" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2605.11814", "pdf_url": "https://arxiv.org/pdf/2605.11814", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.12061", "title": "SAGE: A Self-Evolving Agentic Graph-Memory Engine for Structure-Aware Associative Memory", "abstract": "Long-term memory is becoming a central bottleneck for language agents. Exsting RAG and GraphRAG systems largely treat memory graphs as static retrieval middleware, which limits their ability to recover complete evidence chains from partial cues, exploit reusable graph-structrual roles, and improve the memory itself through downstream feedback. We introduce SAGE, a Self-evolving Agentic Graph-memory Engine that models graph memory as a dynamic long-term memory substrate. SAGE couples two roles: a memory writer that incrementally constucts structured graph memory from interaction histories, and a Graph Foundation Model-based memory reader to perform retrieval and provide feedback to the memory writer. We provide rigorooous theoretical annalyses supporting the framework. Across multi-hop QA, open-domain retireval, domain-specific review QA, and long-term agent-memory benchmarks, SAGE improves evidence recovery, answer grounding, and retrieval efficiency: after two self-evolution rounds, it achieves the best average rank on multi-hop QA; in zero-shot open-domain transfer, it reaches 82.5/91.6 Recall@2/5 on NQ. Further results on LongMemEval and HaluMem show that traning and reader-writer feedback improve multiple long-term memory and hallucination-diagnostic metrics, suggesting that self-evolving, structure-aware graph memory is a promising foundation for robust long-horizon language agents.", "published": "2026-05-12", "updated": "2026-05-12", "authors": [ "Juntong Wang", "Haoyue Zhao", "guanghui Pan", "Xiyuan Wang", "Yanbo Wang", "Qiyan Deng", "Muhan Zhang" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2605.12061", "pdf_url": "https://arxiv.org/pdf/2605.12061", "matched_queries": [ "agent-memory", "long-term-memory", "memory-evaluation" ] }, { "arxiv_id": "2605.12213", "title": "Goal-Oriented Reasoning for RAG-based Memory in Conversational Agentic LLM Systems", "abstract": "LLM-based conversational AI agents struggle to maintain coherent behavior over long horizons due to limited context. While RAG-based approaches are increasingly adopted to overcome this limitation by storing interactions in external memory modules and performing retrieval from them, their effectiveness in answering challenging questions (e.g., multi-hop, commonsense) ultimately depends on the agent's ability to reason over the retrieved information. However, existing methods typically retrieve memory based on semantic similarity to the raw user utterance, which lacks explicit reasoning about missing intermediate facts and often returns evidence that is irrelevant or insufficient for grounded reasoning. In this work, we introduce Goal-Mem, a goal-oriented reasoning framework for RAG-based agentic memory that performs explicit backward chaining from the user's utterance as a goal. Rather than progressively expanding from retrieved context, Goal-Mem decomposes each goal into atomic subgoals, performs targeted memory retrieval to satisfy each subgoal, and iteratively identifies what information from memory should be retrieved when intermediate goals cannot be resolved. We formalize this process in Natural Language Logic, a logical system that combines the verifiability of reasoning provided by FOL with the expressivity of natural language. Through extensive experiments on two datasets and comparing to nine strong memory baselines, we show that Goal-Mem consistently improves performance, particularly on tasks requiring multi-hop reasoning and implicit inference.", "published": "2026-05-12", "updated": "2026-06-07", "authors": [ "Jiazhou Liang", "Armin Toroghi", "Yifan Simon Liu", "Faeze Moradi Kalarde", "Liam Gallagher", "Scott Sanner" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2605.12213", "pdf_url": "https://arxiv.org/pdf/2605.12213", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.12294", "title": "Executable Agentic Memory for GUI Agent", "abstract": "Modern GUI agents typically rely on a model-centric and step-wise interaction paradigm, where LLMs must re-interpret the UI and re-decide actions at every screen, which is fragile in long-horizon tasks. In this paper, we propose Executable Agentic Memory (EAM), a structured Knowledge Graph (KG) that shifts GUI planning from free-form generation to a robust retrieval-and-execution process. Our approach includes a sample-efficient memory construction pipeline using state-aware DFS and action-group mining to compress multi-step routines. To ensure efficient planning, we introduce a value-guided graph search where a lightweight Q-function model steers Monte Carlo Tree Search (MCTS) over the KG. We theoretically establish bias-consistency for the Q-model and derive sample complexity bounds for path recovery. Empirically, EAM outperforms state-of-the-art baselines like UI-TARS-7B by up to $19.6\\%$ on AndroidWorld, while reducing token costs $6\\times$ relative to GPT-4o. With a $2.8$s average latency, EAM enables reliable, quick, and long-horizon GUI automation.", "published": "2026-05-12", "updated": "2026-05-12", "authors": [ "Zerui Qin", "Sheng Yue", "Xingyuan Hua", "Yongjian Fu", "Ju Ren" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2605.12294", "pdf_url": "https://arxiv.org/pdf/2605.12294", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.12493", "title": "LongMemEval-V2: Evaluating Long-Term Agent Memory Toward Experienced Colleagues", "abstract": "Long-term memory is crucial for agents in specialized web environments, where success depends on recalling interface affordances, state dynamics, workflows, and recurring failure modes. However, existing memory benchmarks for agents mostly focus on user histories, short traces, or downstream task success, leaving open how to directly evaluate whether memory systems effectively internalize environment-specific experience. To address this gap, we introduce LongMemEval-V2 (LME-V2), a benchmark for evaluating whether memory systems can help agents acquire the experience needed to become knowledgeable colleagues in customized environments. LME-V2 contains 451 manually curated questions covering five core memory abilities for web agents: static state recall, dynamic state tracking, workflow knowledge, environment gotchas, and premise awareness. Questions are paired with history trajectories containing up to 500 trajectories and 115M tokens. We use a context gathering formulation: memory systems consume history trajectories and return compact evidence for downstream question answering. We propose a suite of two memory methods: AgentRunbook-R, an efficient RAG-based memory with knowledge pools for raw state observations, events, and strategy notes, and AgentRunbook-C, which stores trajectories as files and invokes a coding agent to gather evidence in an augmented sandbox. Experiments show that AgentRunbook-C achieves the best performance with 72.5% average accuracy, outperforming the strongest RAG baseline (48.5%) and the off-the-shelf coding agent baseline (69.3%). Despite the strong performance gains, coding agent based methods have high latency costs. While AgentRunbook-C advances the accuracy-latency Pareto frontier, substantial room for improvement remains. Together, these results establish LME-V2 as a challenging testbed for developing long-term memory systems for environment experience.", "published": "2026-05-12", "updated": "2026-05-12", "authors": [ "Di Wu", "Zixiang Ji", "Asmi Kawatkar", "Bryan Kwan", "Jia-Chen Gu", "Nanyun Peng", "Kai-Wei Chang" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2605.12493", "pdf_url": "https://arxiv.org/pdf/2605.12493", "matched_queries": [ "agent-memory", "memory-evaluation" ] }, { "arxiv_id": "2605.12978", "title": "Useful Memories Become Faulty When Continuously Updated by LLMs", "abstract": "Learning from past experience benefits from two complementary forms of memory: episodic traces -- raw trajectories of what happened -- and consolidated abstractions distilled across many episodes into reusable, schema-like lessons. Recent agentic-memory systems pursue the consolidated form: an LLM rewrites past trajectories into a textual memory bank that it continuously updates with new interactions, promising self-improving agents without parameter updates. Yet we find that such consolidated memories produced by today's LLMs are often faulty even when derived from useful experiences. As consolidation proceeds, memory utility first rises, then degrades, and can fall below the no-memory baseline. More surprisingly, even when consolidating from ground-truth solutions, GPT-5.4 fails on 54% of a set of ARC-AGI problems it had previously solved without memory. We trace the regression to the consolidation step rather than the underlying experience: the same trajectories yield qualitatively different memories under different update schedules, and an episodic-only control that simply retains those trajectories remains competitive with the consolidators we test. In a controlled ARC-AGI Stream environment that exposes Retain, Delete, and Consolidate actions, agents preserve raw episodes by default and double the accuracy of their forced-consolidation counterparts; disabling consolidation entirely (episodic management only) matches this auto regime. Practically, robust agent memory should treat raw episodes as first-class evidence and gate consolidation explicitly rather than firing it after every interaction. Looking forward, reliable agentic memory will require LLMs that can consolidate without overwriting the evidence they depend on.", "published": "2026-05-13", "updated": "2026-05-13", "authors": [ "Dylan Zhang", "Yanshan Lin", "Zhengkun Wu", "Yihang Sun", "Bingxuan Li", "Dianqi Li", "Hao Peng" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2605.12978", "pdf_url": "https://arxiv.org/pdf/2605.12978", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.13438", "title": "CogniFold: Always-On Proactive Memory via Cognitive Folding", "abstract": "Existing agent memory remains predominantly reactive and retrieval-based, lacking the capacity to autonomously organize experience into persistent cognitive structure. Toward genuinely autonomous agents, we introduce CogniFold, a brain-inspired \"always-on\" agent memory designed for the next generation of proactive assistants. CogniFold continuously folds fragmented event streams into self-emerging cognitive structures, bootstrapping progressively higher-level cognition from incoming events and accumulated knowledge. We ground this by extending Complementary Learning Systems (CLS) theory from two layers (hippocampus, neocortex) to three, adding a prefrontal intent layer. Emulating the prefrontal cortex as the locus of intentional control and decision-making, CogniFold achieves this through graph-topology self-organization: cognitive structures proactively assemble under the stream, merge when semantically similar, decay when stale, relink through associative recall, and surface intents when concept-cluster density crosses a threshold. We evaluate structural formation using CogEval-Bench, demonstrating that CogniFold uniquely produces memory structures that match cognitive expectations and concept emergence. Furthermore, across eight downstream benchmarks -- two probing long-term conversational memory (LoCoMo, LongMemEval) and six spanning other cognitive domains -- we validate that CogniFold simultaneously performs robustly on conventional memory tasks. Our code is available at https://github.com/OpenNorve/CogniFold.", "published": "2026-05-13", "updated": "2026-06-17", "authors": [ "Suli Wang", "Yiqun Duan", "Yu Deng", "Rundong Zhao", "Dai Shi", "Minghua Deng", "Chen Chen", "Xinliang Zhou" ], "categories": [ "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2605.13438", "pdf_url": "https://arxiv.org/pdf/2605.13438", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.13542", "title": "RealICU: Do LLM Agents Understand Long-Context ICU Data? A Benchmark Beyond Behavior Imitation", "abstract": "Intensive care units (ICU) generate long, dense and evolving streams of clinical information, where physicians must repeatedly reassess patient states under time pressure, underscoring a clear need for reliable AI decision support. Existing ICU benchmarks typically treat historical clinician actions as ground truth. However, these actions are made under incomplete information and limited temporal context of the underlying patient state, and may therefore be suboptimal, making it difficult to assess the true reasoning capabilities of AI systems. We introduce RealICU, a hindsight-annotated benchmark for evaluating large language models (LLMs) under realistic ICU conditions, where labels are created after senior physicians review the full patient trajectory. We formulate four physician-motivated tasks: assess Patient Status, Acute Problems, Recommended Actions, and Red Flag actions that risk unsafe outcomes. We partition each trajectory with 30-min windows and release two datasets: RealICU-Gold with 930-window annotations from 94 MIMIC-IV patients, and RealICU-Scale with 11,862 windows extended by Oracle, a physician-validated LLM hindsight labeler. Existing LLMs including memory-augmented ones performed poorly on RealICU, exposing two failure modes: a recall-safety tradeoff for clinical recommendations, and an anchoring bias to early interpretations of the patient. We further introduce ICU-Evo to study structured-memory agents that improves long-horizon reasoning but does not fully eliminate safety failures. Together, RealICU provides a clinically grounded testbed for measuring and improving AI sequential decision-support in high-stakes care. Project page: https://chengzhi-leo.github.io/RealICU-Bench/", "published": "2026-05-13", "updated": "2026-05-13", "authors": [ "Chengzhi Shen", "Weixiang Shen", "Tobias Susetzky", "Chen", "Chen", "Jun Li", "Yuyuan Liu", "Xuepeng Zhang", "Zhenyu Gong", "Daniel Rueckert", "Jiazhen Pan" ], "categories": [ "cs.AI", "cs.CL", "cs.LG", "cs.MA" ], "url": "https://arxiv.org/abs/2605.13542", "pdf_url": "https://arxiv.org/pdf/2605.13542", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.13941", "title": "EvolveMem:Self-Evolving Memory Architecture via AutoResearch for LLM Agents", "abstract": "Long-term memory is essential for LLM agents that operate across multiple sessions, yet existing memory systems treat retrieval infrastructure as fixed: stored content evolves while scoring functions, fusion strategies, and answer-generation policies remain frozen at deployment. We argue that truly adaptive memory requires co-evolution at two levels: the stored knowledge and the retrieval mechanism that queries it. We present EvolveMem, a self-evolving memory architecture that exposes its full retrieval configuration as a structured action space optimized by an LLM-powered diagnosis module. In each evolution round, the module reads per-question failure logs, identifies root causes, and proposes targeted configuration adjustments; a guarded meta-analyzer applies them with automatic revert-on-regression and explore-on-stagnation safeguards. This closed-loop self-evolution realizes an AutoResearch process: the system autonomously conducts iterative research cycles on its own architecture, replacing manual configuration tuning. Starting from a minimal baseline, the process converges autonomously, discovering effective retrieval strategies including entirely new configuration dimensions not present in the original action space. On LoCoMo, EvolveMem outperforms the strongest baseline by 25.7% relative and achieves a 78.0% relative improvement over the minimal baseline. On MemBench, EvolveMem exceeds the strongest baseline by 18.9% relative. Evolved configurations transfer across benchmarks with positive rather than catastrophic transfer, indicating that the self-evolution process captures universal retrieval principles rather than benchmark-specific heuristics. Code is available at https://github.com/aiming-lab/SimpleMem.", "published": "2026-05-13", "updated": "2026-05-13", "authors": [ "Jiaqi Liu", "Xinyu Ye", "Peng Xia", "Zeyu Zheng", "Cihang Xie", "Mingyu Ding", "Huaxiu Yao" ], "categories": [ "cs.LG", "cs.AI" ], "url": "https://arxiv.org/abs/2605.13941", "pdf_url": "https://arxiv.org/pdf/2605.13941", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2605.14421", "title": "MemLineage: Lineage-Guided Enforcement for LLM Agent Memory", "abstract": "We introduce MemLineage, a defense for LLM agent memory that attaches both cryptographic provenance and LLM-mediated derivation lineage to every entry. Recent and concurrent work shows that untrusted content can be written into persistent agent state and re-enter later sessions as an instruction; the remaining systems question is how to preserve useful memory recall while preventing such state from justifying sensitive actions. MemLineage treats this as a chain-of-custody problem rather than a filtering problem. It is a six-module design around an RFC-6962 Merkle log over per-principal Ed25519-signed entries: a weighted derivation DAG records which retrieved entries influenced each new memory, and a max-of-strong-edges propagation rule makes Untrusted-Path Persistence hold for any chain whose attribution edges remain above threshold. The sensitive-action gate then refuses dispatches whose active justification descends from an external ancestor, while still allowing benign recall. We evaluate three defense cells against three memory-poisoning workloads on a deterministic mechanism-isolation harness; MemLineage is the only configuration in that harness that drives all three columns to zero ASR, while sub-millisecond per-operation overhead keeps it well below the noise floor of any LLM call. A Codex-backed AgentDojo bridge further separates strong-model behavior from defense-layer behavior: under an intentionally vulnerable tool-output profile, no-defense and signature-only baselines fail on all six banking pairs, while all MemLineage rows reduce strict AgentDojo ASR to zero. The core deterministic artifacts are byte-equal CI-verified; hosted-model AgentDojo and live-model sweeps are recorded as auditable logs rather than byte-pinned artifacts.", "published": "2026-05-14", "updated": "2026-05-14", "authors": [ "Ciyan Ouyang", "Rui Hou" ], "categories": [ "cs.CR", "cs.AI" ], "url": "https://arxiv.org/abs/2605.14421", "pdf_url": "https://arxiv.org/pdf/2605.14421", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.14498", "title": "GroupMemBench: Benchmarking LLM Agent Memory in Multi-Party Conversations", "abstract": "Large Language Model (LLM) agents increasingly serve as personal assistants and workplace collaborators, where their utility depends on memory systems that extract, retrieve, and apply information across long-running conversations. However, both existing memory systems and benchmarks are built around the dyadic, single-user setup, even though real deployments routinely span groups and channels with multiple users interacting with the agent and with each other. This mismatch leaves three properties of group memory unmeasured: (i) group dynamics that go beyond concatenated one-on-one chats, (ii) speaker-grounded belief tracking, where the per-user memory modeling is needed, and (iii) audience-adapted language, where Theory-of-Mind shifts produce role-specific vocabulary. We introduce GroupMemBench, a benchmark that exposes all three. A graph-grounded synthesis pipeline produces multi-party conversations with controllable reply structure and conditions each message on per-user personas and target audiences. An adversarial query pipeline then binds every question to a specific asker across six categories, spanning multi-hop reasoning, knowledge update, term ambiguity, user-implicit reasoning, temporal reasoning, and abstention, and iteratively searches challenging, realistic queries that reflect comprehensive memory capability. Benchmarking leading memory systems exposes a sharp collapse: the strongest one reaches only 46.0% average accuracy, with knowledge update at 27.1% and term ambiguity at 37.7%, while a simple BM25 baseline matches or exceeds most agent memory systems. This indicates current memory ingestion erases the structural and lexical features group memory depends on, leaving multi-user memory far from solved.", "published": "2026-05-14", "updated": "2026-05-16", "authors": [ "Jingbo Yang", "Kwei-Herng Lai", "Xiaowen Wang", "Shiyu Chang", "Yaar Harari", "Evgeniy Gabrilovich" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2605.14498", "pdf_url": "https://arxiv.org/pdf/2605.14498", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.14906", "title": "MemLens: Benchmarking Multimodal Long-Term Memory in Large Vision-Language Models", "abstract": "Memory is essential for large vision-language models (LVLMs) to handle long, multimodal interactions, with two method directions providing this capability: long-context LVLMs and memory-augmented agents. However, no existing benchmark conducts a systematic comparison of the two on questions that genuinely require multimodal evidence. To close this gap, we introduce MEMLENS, a comprehensive benchmark for memory in multimodal multi-session conversations, comprising 789 questions across five memory abilities (information extraction, multi-session reasoning, temporal reasoning, knowledge update, and answer refusal) at four standard context lengths (32K-256K tokens) under a cross-modal token-counting scheme. An image-ablation study confirms that solving MEMLENS requires visual evidence: removing evidence images drops two frontier LVLMs below 2% accuracy on the 80.4% of questions whose evidence includes images. Evaluating 27 LVLMs and 7 memory-augmented agents, we find that long-context LVLMs achieve high short-context accuracy through direct visual grounding but degrade as conversations grow, whereas memory agents are length-stable but lose visual fidelity under storage-time compression. Multi-session reasoning caps most systems below 30%, and neither approach alone solves the task. These results motivate hybrid architectures that combine long-context attention with structured multimodal retrieval. Our code is available at https://github.com/xrenaf/MEMLENS.", "published": "2026-05-14", "updated": "2026-05-14", "authors": [ "Xiyu Ren", "Zhaowei Wang", "Yiming Du", "Zhongwei Xie", "Chi Liu", "Xinlin Yang", "Haoyue Feng", "Wenjun Pan", "Tianshi Zheng", "Baixuan Xu", "Zhengnan Li", "Yangqiu Song", "Ginny Wong", "Simon See" ], "categories": [ "cs.CV" ], "url": "https://arxiv.org/abs/2605.14906", "pdf_url": "https://arxiv.org/pdf/2605.14906", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.15128", "title": "MemEye: A Visual-Centric Evaluation Framework for Multimodal Agent Memory", "abstract": "Long-term agent memory is increasingly multimodal, yet existing evaluations rarely test whether agents preserve the visual evidence needed for later reasoning. In prior work, many visually grounded questions can be answered using only captions or textual traces, allowing answers to be inferred without preserving the fine-grained visual evidence. Meanwhile, harder cases that require reasoning over changing visual states are largely absent. Therefore, we introduce MemEye, a framework that evaluates memory capabilities from two dimensions: one measures the granularity of decisive visual evidence (from scene-level to pixel-level evidence), and the other measures how retrieved evidence must be used (from single evidence to evolutionary synthesis). Under this framework, we construct a new benchmark across 8 life-scenario tasks, with ablation-driven validation gates for assessing answerability, shortcut resistance, visual necessity, and reasoning structure. By evaluating 13 memory methods across 4 VLM backbones, we show that current architectures still struggle to preserve fine-grained visual details and reason about state changes over time. Our findings show that long-term multimodal memory depends on evidence routing, temporal tracking, and detail extraction.", "published": "2026-05-14", "updated": "2026-05-14", "authors": [ "Minghao Guo", "Qingyue Jiao", "Zeru Shi", "Yihao Quan", "Boxuan Zhang", "Danrui Li", "Liwei Che", "Wujiang Xu", "Shilong Liu", "Zirui Liu", "Mubbasir Kapadia", "Vladimir Pavlovic", "Jiang Liu", "Mengdi Wang", "Yiyu Shi", "Dimitris N. Metaxas", "Ruixiang Tang" ], "categories": [ "cs.CV", "cs.CL", "cs.IR" ], "url": "https://arxiv.org/abs/2605.15128", "pdf_url": "https://arxiv.org/pdf/2605.15128", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.15338", "title": "Hidden in Memory: Sleeper Memory Poisoning in LLM Agents", "abstract": "Large language models are increasingly augmented with persistent memory, allowing assistants to store user-specific information across sessions for personalization and continuity. This statefulness introduces a new security risk: adversarial content can corrupt what an assistant remembers and thereby influence future interactions. We propose and study sleeper memory poisoning, a delayed attack in which an adversary manipulates external context, such as a document, webpage, or repository, to cause the assistant to store a fabricated memory about the user. Unlike conventional prompt injection, the attack can remain dormant and re-emerge across multiple later conversations. We evaluate the full attack pipeline: whether poisoned memories are written, later retrieved, and ultimately used to steer the following conversations. Across stateful LLM assistants, poisoned memories were added up to 99.8% on GPT-5.5 and 95% on Kimi-K2.6. Crucially, among successful retrievals, poisoned memories cause attacker-intended agentic actions in 60-89% of evaluations across models. These results show that persistent memory can act as a long-term attack surface across multiple future conversations.", "published": "2026-05-14", "updated": "2026-05-18", "authors": [ "Sidharth Pulipaka", "Stanislau Hlebik", "Leonidas Raghav", "Sahar Abdelnabi", "Vyas Raina", "Ivaxi Sheth", "Mario Fritz" ], "categories": [ "cs.CR", "cs.AI" ], "url": "https://arxiv.org/abs/2605.15338", "pdf_url": "https://arxiv.org/pdf/2605.15338", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2605.15701", "title": "H-Mem: A Novel Memory Mechanism for Evolving and Retrieving Agent Memory via a Hybrid Structure", "abstract": "Memory data are ubiquitous in Large Language Model (LLM)-based agents (e.g., OpenClaw and Manus). A few recent works have attempted to exploit agents'memory for improving their performance on the question-answering (QA) task, but they lack a principled mechanism for effectively modeling how memory data evolves over time and retrieving memory data effectively, leading to poor performance in memory utilization. To fill this gap, we present H-Mem, a novel memory mechanism via a hybrid structure that can not only effectively model the evolution of agent memory over a long period of time, but also provide an efficient memory retrieval approach. Particularly, H-Mem builds a temporal and semantic tree structure that allows the short-term memory data to evolve progressively into long-term memory data, where the latter provides summarized information about the former, while simultaneously constructing a knowledge graph to capture the relationships between entities in memory. Moreover, it offers an effective memory retrieval approach by exploiting the hybrid structure of the tree and graph structures. Extensive experiments on three agent memory benchmarks show that H-Mem achieves state-of-the-art performance on the QA task.", "published": "2026-05-15", "updated": "2026-05-15", "authors": [ "Jiawei Yu", "Yixiang Fang", "Xilin Liu", "Yuchi Ma" ], "categories": [ "cs.CL", "cs.AI" ], "url": "https://arxiv.org/abs/2605.15701", "pdf_url": "https://arxiv.org/pdf/2605.15701", "matched_queries": [ "agent-memory", "memory-evaluation" ] }, { "arxiv_id": "2605.15710", "title": "SMMBench: A Benchmark for Source-Distributed Multimodal Agent Memory", "abstract": "Existing benchmarks for multimodal memory reasoning largely evaluate systems within pre-assembled contexts, but under-evaluate whether agents can use evidence distributed across independently originated sources. We argue that source-distributed memory composition is an important and under-examined bottleneck in multimodal agent memory, especially when relevant evidence is fragmented across heterogeneous artifacts such as conversations, profiles, screenshots, tables, images, and documents. To address this gap, we introduce Source-distributed Multimodal Memory Benchmark(SMMBench), which measures whether agents can retrieve, align, and compose multimodal evidence scattered across multiple sources rather than reason within a single curated context. SMMBench evaluates four core capabilities: (1) cross-source multimodal reasoning; (2) conflict resolution; (3) preference reasoning; (4) memory-grounded action prediction. The benchmark contains 1877 samples grounded in 264 sources. Experiments on representative memory-style and retrieval-based baselines show that current systems still struggle on these capabilities, positioning source-distributed multimodal memory as an important and still under-evaluated challenge for multimodal agents. Our data are available at https://huggingface.co/datasets/HuacanChai/SMMBench.", "published": "2026-05-15", "updated": "2026-05-15", "authors": [ "Huacan Chai", "Yukai Wang", "Yingxuan Yang", "Dan Peng", "Yuanyi Song", "Zhihui Fu", "Weiwen Liu", "Jianghao Lin", "Jun Wang", "Weinan Zhang" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2605.15710", "pdf_url": "https://arxiv.org/pdf/2605.15710", "matched_queries": [ "agent-memory", "memory-evaluation" ] }, { "arxiv_id": "2605.15759", "title": "DimMem: Dimensional Structuring for Efficient Long-Term Agent Memory", "abstract": "Large language model (LLM) agents require long-term memory to leverage information from past interactions. However, existing memory systems often face a fidelity--efficiency trade-off: raw dialogue histories are expensive, while flat facts or summaries may discard the structure needed for precise recall. We propose \\textbf{DimMem}, a lightweight dimensional memory framework that represents each memory as an atomic, typed, and self-contained unit with explicit fields such as time, location, reason, purpose, and keywords. This representation exposes the structure needed for dimension-aware retrieval, memory update, and selective assistant-context recall without storing full histories in the model context. Across LoCoMo-10 and LongMemEval-S, DimMem achieves \\textbf{81.43\\%} and \\textbf{78.20\\%} overall accuracy, respectively, outperforming existing lightweight memory systems while reducing LoCoMo per-query token cost by \\textbf{24\\%}. We further show that dimensional memory extraction is learnable by compact models: after fine-tuning on the DimMem schema, a Qwen3-4B extractor surpasses LightMem with GPT-4.1-mini on both benchmarks and reaches performance comparable to, or better than, much larger extractors in key settings. These results suggest that explicit dimensional structuring is an effective and efficient foundation for long-term memory in LLM agents. Code is available at https://github.com/ChowRunFa/DimMem.", "published": "2026-05-15", "updated": "2026-05-24", "authors": [ "Wentao Qiu", "Haotian Hu", "Fanyi Wang", "Jinwei Kong", "Yu Zhang" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2605.15759", "pdf_url": "https://arxiv.org/pdf/2605.15759", "matched_queries": [ "agent-memory", "long-term-memory" ] }, { "arxiv_id": "2605.16233", "title": "FORGE: Self-Evolving Agent Memory With No Weight Updates via Population Broadcast", "abstract": "Can LLM agents improve decision-making through self-generated memory without gradient updates? We propose FORGE (Failure-Optimized Reflective Graduation and Evolution), a staged, population-based protocol that evolves prompt-injected natural-language memory for hierarchical ReAct agents. FORGE wraps a Reflexion-style inner loop, where a dedicated reflection agent (using the same underlying LLM, no distillation from a stronger model) converts failed trajectories into reusable knowledge artifacts: textual heuristics (Rules), few-shot demonstrations (Examples), or both (Mixed), with an outer loop that propagates the best-performing instance's memory to the population between stages and freezes converged instances via a graduation criterion. We evaluate on CybORG CAGE-2, a stochastic network-defense POMDP at a 30-step horizon against the B-line attacker, where all four tested LLM families (Gemini-2.5-Flash-Lite, Grok-4-Fast, Llama-4-Maverick, Qwen3-235B) exhibit strongly negative, heavy-tailed zero-shot rewards. Compared against both a zero-shot baseline and a Reflexion baseline (isolated single-stream learning), FORGE improves average evaluation return by 1.7-7.7$\\times$ over zero-shot and by 29-72% over Reflexion in all 12 model-representation conditions, reducing major-failure rates (below $-100$) to as low as $\\sim$1%. We find that (1) population broadcast is critical mechanism, with a no-graduation ablation confirming that broadcast carries the performance gains while graduation primarily saves compute; (2) Examples achieves the strongest returns for three of four models, Rules offers the best cost-reliability profile with $\\sim$40% fewer tokens; and (3) weaker baseline models benefit disproportionately, suggesting FORGE may mitigate capability gaps rather than amplify strong models. All evidence is confined to CAGE-2 B-line; cross-family findings are directional evidence.", "published": "2026-05-15", "updated": "2026-05-15", "authors": [ "Igor Bogdanov", "Chung-Horng Lung", "Thomas Kunz", "Jie Gao", "Adrian Taylor", "Marzia Zaman" ], "categories": [ "cs.AI", "cs.CL", "cs.LG", "cs.MA", "eess.SY" ], "url": "https://arxiv.org/abs/2605.16233", "pdf_url": "https://arxiv.org/pdf/2605.16233", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.16481", "title": "Visual Agentic Memory: Enabling Online Long Video Understanding via Online Indexing, Hierarchical Memory, and Agentic Retrieval", "abstract": "Long video understanding requires more than large context windows. It also needs a memory mechanism that decides what visual evidence to retain, keeps it searchable over long horizons, and grounds later reasoning in recoverable observations rather than compressed latent state alone. We propose Visual Agentic Memory (VAM), a training-free framework with three components. Online Indexing supports selective evidence retention under streaming constraints. Hierarchical Memory organises retained evidence in a Parallel Representation that aligns temporal context with spatial observations. Agentic Retrieval searches, inspects, and verifies candidate evidence before producing a grounded answer. On OVO-Bench, VAM achieves the highest RT+BT average (68.41) across all reported baselines, improving over end-to-end use of the same underlying MLLM (Gemini 3 Flash, 67.46). On the month-scale split of MM-Lifelong train@month (105.6 hours over 51 days), VAM reaches 17.11%, second only to ReMA with GPT-5 (17.62%). These results suggest that long-horizon video understanding benefits from treating visual memory as an explicit, inspectable, and queryable substrate. Code is available at https://github.com/yiliu-li/Visual-Agentic-Memory.", "published": "2026-05-15", "updated": "2026-05-15", "authors": [ "Aiden Yiliu Li", "Nels Numan", "Anthony Steed" ], "categories": [ "cs.CV", "cs.AI" ], "url": "https://arxiv.org/abs/2605.16481", "pdf_url": "https://arxiv.org/pdf/2605.16481", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.16746", "title": "State Contamination in Memory-Augmented LLM Agents", "abstract": "LLM agents increasingly rely on persistent state, including transcripts, summaries, retrieved context, and memory buffers, to support long-horizon interaction. This makes safety depend not only on individual model outputs, but also on what an agent stores and later reuses. We study a failure mode we call memory laundering: toxic or adversarial context can be compressed into memory summaries that no longer appear toxic under standard detectors, while still preserving hostile framing or conflict structure that influences future generations. Using paired counterfactual multi-agent rollouts, we show that toxic-origin memory summaries can remain below common toxicity thresholds while nevertheless increasing downstream toxicity relative to matched neutral baselines. To measure this hidden influence, we introduce the sub-threshold propagation gap (SPG), which quantifies downstream behavioral differences conditioned on memory states that a deployed monitor would classify as safe. Our experiments show that toxicity propagates through distinct state channels: raw transcript reuse drives overt downstream toxicity, while compressed memory carries hidden sub-threshold influence. We further find that mitigation depends critically on intervention placement. Sanitizing toxic state before summarization substantially reduces the hidden propagation gap, whereas cleaning only the completed summary can leave laundered influence intact. These results suggest that safety in memory-augmented agents should be treated as a state-control problem over evolving context, with sanitization applied before unsafe information is compressed into persistent memory.", "published": "2026-05-16", "updated": "2026-05-16", "authors": [ "Yian Wang", "Agam Goyal", "Yuen Chen", "Hari Sundaram" ], "categories": [ "cs.AI", "cs.LG" ], "url": "https://arxiv.org/abs/2605.16746", "pdf_url": "https://arxiv.org/pdf/2605.16746", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2605.16858", "title": "Pedestrian-Aware LLM-Driven Behavioral Planning for Autonomous Vehicles", "abstract": "Autonomous Vehicles (AVs) must make reliable decisions in dense urban environments where pedestrian behavior is variable, sometimes abnormal, and often unseen during training. Reinforcement learning (RL)-based AV control systems perform well in structured traffic but struggle to generalize to unpredictable pedestrian interactions and out-of-distribution scenarios. Their reliance on handcrafted rewards and opaque decisions further limits their suitability for safety-critical, pedestrian-rich environments. To address these limitations, we introduce a Large Language Model (LLM)-based decision-making framework for pedestrian-aware behavioral planning. The system converts structured scene observations into natural-language reasoning prompts, enabling the LLM to infer pedestrian intent, anticipate risk, and generate cautious tactical driving decisions. These decisions are executed by a motion planner that ensures smooth, kinematically feasible control. We evaluate the framework in SUMO across multiple pedestrian-interaction scenarios, including unexpected jaywalking, turn-back crossing, hesitation, and bidirectional crossing. In zero-shot evaluation, the LLM-based agent achieves a 68% collision-free success rate, substantially outperforming deep RL baselines (17.7%). With few-shot episodic memory in a single-pedestrian scenario, performance increases to 96.0%, exceeding a custom DQN controller (82.0%). Cross-behavior evaluation further shows that memory derived from turn-back interactions transfers to unseen hesitation and bidirectional crossing scenarios, achieving 82.0% and 90.0% success, respectively. The system consistently initiates earlier responses, maintains wider safety buffers, and produces interpretable, human-aligned decisions.", "published": "2026-05-16", "updated": "2026-05-16", "authors": [ "Aidana Baimbetova", "Haruki Yonekura", "Hamada Rizk", "Hirozumi Yamaguchi" ], "categories": [ "cs.RO", "cs.AI" ], "url": "https://arxiv.org/abs/2605.16858", "pdf_url": "https://arxiv.org/pdf/2605.16858", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2605.23986", "title": "MemForest: An Efficient Agent Memory System with Hierarchical Temporal Indexing", "abstract": "Memory is a fundamental component for enabling long-context LLM agents, supporting persistent state across interactions through a continuous serve-and-update lifecycle. Despite substantial prior work, existing systems suffer from significant maintenance overhead due to two key limitations: coarse-grained state management and inherently sequential update pipelines. In particular, updates are often tightly coupled with LLM inference and require full-state rewrites, leading to poor scalability and growing latency as memory accumulates. To address these challenges, we present MemForest, a memory framework that reformulates agent memory as a write-efficient temporal data management problem. MemForest breaks the sequential bottleneck via parallel chunk extraction, decoupling memory construction into concurrent, independent operations. To further eliminate coarse-grained maintenance, we introduce MemTree, a hierarchical temporal index that organizes memory as time-ordered trees rather than flat global summaries. This design replaces full-state rewrites with localized per-node updates, reducing maintenance cost to the affected tree paths while naturally preserving temporally evolving states. We evaluate MemForest on two long-context memory benchmarks, LongMemEval-S and LoCoMo. On LongMemEval-S, MemForest achieves the best overall performance among stateful baselines, reaching 79.8% pass@1 accuracy while sustaining a memory construction throughput approximately 6x higher than state-of-the-art approaches including EverMemOS.", "published": "2026-05-16", "updated": "2026-05-16", "authors": [ "Han Chen", "Zining Zhang", "Wenqi Pei", "Bingsheng He", "Ming Wu", "Jason Zeng", "Michael Heinrich", "Wei Wu", "Hongbao Zhang" ], "categories": [ "cs.DB", "cs.AI", "cs.MA" ], "url": "https://arxiv.org/abs/2605.23986", "pdf_url": "https://arxiv.org/pdf/2605.23986", "matched_queries": [ "agent-memory", "memory-evaluation" ] }, { "arxiv_id": "2605.17348", "title": "Taming \"Zombie'' Agents: A Markov State-Aware Framework for Resilient Multi-Agent Evolution", "abstract": "Recent advancements in LLM-based multi-agent systems have demonstrated remarkable collaborative capabilities across complex tasks. To improve overall efficiency, existing methods often rely on aggressive graph evolution among agents (e.g., node or edge pruning), which risks prematurely discarding valuable agents due to transient issues such as hallucinations or temporary knowledge gaps. However, such hard pruning overlooks the potential for ``zombie'' agents to recover and contribute in subsequent discussion rounds. In this paper, we propose AgentRevive, a Markov state-aware framework for resilient multi-agent evolution. Our approach dynamically manages agent collaboration through soft state transitions, implemented via two key components: (1) State-Aware Policy Learning: Agent states are divided into ``Active'', ``Standby'', and ``Terminated'' states, selectively propagating messages based on agent memory. The policy employs a risk estimator to optimize agent state transitions by assessing hallucination risk, minimizing the influence of unreliable nodes while safeguarding valuable ones. (2) State-Aware Edge Optimization: Subgraph edges are pruned according to states learned from the policy, permanently removing ``Terminated'' nodes and retaining ``Standby'' nodes for subsequent rounds to assess their potential future contributions. Extensive experiments on general reasoning, domain-specific, and hallucination challenge tasks show that our method consistently outperforms strong baselines and significantly reduces token consumption through state-aware agent scheduling.", "published": "2026-05-17", "updated": "2026-05-17", "authors": [ "Taolin Zhang", "Pukun Zhao", "Qizhou Chen", "Jiuheng Wan", "Chen Chen", "Xiaofeng He", "Chengyu Wang", "Richang Hong" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2605.17348", "pdf_url": "https://arxiv.org/pdf/2605.17348", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.17596", "title": "NeuSymMS: A Hybrid Neuro-Symbolic Memory System for Persistent, Self-Curating LLM Agents", "abstract": "We present NeuSymMS, an adaptive memory system that enables large language model (LLM) agents to learn, remember, and reason about users across sessions via a hybrid neuro-symbolic architecture. NeuSymMS couples neural fact extraction from unstructured dialogue using LLMs and a CLIPS-based expert system that classifies, deduplicates, and reconciles facts under explicit lifecycle rules. The system represents knowledge as subject-relation-value triples stored in relational database management system. It supports user/agents/agent-to-agent scoping, and implements a dual-horizon (short-term and long-term) memory model. IT leverages access-based promotion and time-based pruning of the memory on both horizpons. NeuSymMS maintains continuity of memory while avoiding context-window bloat and cross-entity contamination. We argue that this architecture offers a practical path to trustworthy, auditable memory for production agentic systems and discuss its novelty relative to log retrieval, summarization, and key-value approaches.", "published": "2026-05-17", "updated": "2026-05-20", "authors": [ "Mujahid Sultan", "Sri Thuraisamy", "Daya Rajaratnam" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2605.17596", "pdf_url": "https://arxiv.org/pdf/2605.17596", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2605.17625", "title": "Episodic-Semantic Memory Architecture for Long-Horizon Scientific Agents", "abstract": "As Large Language Models (LLMs) evolve into persistent scientific collaborators, context window saturation has emerged as a critical bottleneck. Scientific workflows involving iterative data analysis and hypothesis refinement rapidly saturate even extended contexts with dense technical content, while monolithic approaches suffer from quadratic cost scaling and cognitive degradation. We evaluate a Dual Process Memory Architecture that decouples immediate episodic needs (constant 10-message window) from long-term consolidated knowledge (growing at approximately 3 tokens/message). Unlike prior social agent memory systems, our domain-specific consolidation addresses contradictory parameter evolution, multi-hop reasoning across experimental phases, and precise technical fact retention. Through large-scale evaluation spanning 15,000 messages with cross-model validation across six LLMs from three families (OpenAI, Anthropic, Google), totaling 1,440 queries, we establish three key findings. First, while full-context models fail at 10,000 messages due to context overflow, our system maintains 70-85% accuracy with 1-2 second latency using 62% fewer tokens (45,434 vs 120,000+ limit). Second, cross-model validation reveals architecture-level trade-offs independent of specific LLMs: Dual Process excels at numeric/temporal queries (65-90% accuracy) while RAG excels at historical retrieval (60-85%), suggesting complementary deployment strategies. Third, we identify a \"Sim-to-Real\" gap where synthetic tests maintain constant memory but realistic workflows exhibit linear growth (about 3 tokens/message), with consolidation quality emerging as the primary scalability bottleneck. The architecture successfully manages profiles with 14,000+ scientific facts (125k tokens), demonstrating that domain-specific memory consolidation enables sustained operation beyond full-context limits.", "published": "2026-05-17", "updated": "2026-05-17", "authors": [ "Nikola Milosevic" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2605.17625", "pdf_url": "https://arxiv.org/pdf/2605.17625", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.17641", "title": "Causal Intervention-Based Memory Selection for Long-Horizon LLM Agents", "abstract": "Long-horizon LLM agents rely on persistent memory to support interactions across sessions, yet existing memory systems often retrieve context using semantic similarity or broad history inclusion, treating retrieved memories as uniformly useful. This assumption is fragile because memories may be topically related while remaining irrelevant, stale, or misleading. We propose Causal Memory Intervention (CMI), a causal memory-selection technique that estimates how candidate memories affect the model's answer under controlled interventions, selecting memories that improve task performance while suppressing unstable, irrelevant, or harmful ones. To evaluate this setting, we introduce Causal-LoCoMo, a causally annotated benchmark derived from long conversational data, where each example contains a user request, a structured memory bank, useful memories, irrelevant distractors, and synthetic harmful memories. We compare CMI against vector, graph, reflection, summary, full-history, and no-memory baselines. Results show that CMI achieves a stronger balance between answer quality and robustness to misleading memory, suggesting that reliable long-term memory requires selecting context based on causal usefulness rather than relevance alone. The full framework, benchmark construction code, and experimental pipeline are available at https://github.com/Saksham4796/causal-memory-intervention.", "published": "2026-05-17", "updated": "2026-05-17", "authors": [ "Saksham Sahai Srivastava" ], "categories": [ "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2605.17641", "pdf_url": "https://arxiv.org/pdf/2605.17641", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2605.18284", "title": "CommitDistill: A Lightweight Knowledge-Centric Memory Layer for Software Repositories", "abstract": "Software repositories accumulate large amounts of unstructured knowledge in commit messages, pull-request discussions, and issue threads, but developers and AI coding assistants rarely reuse this history effectively. Recent work on typed-memory architectures for LLM agents (MemGPT, generative agents, and the PlugMem module of Yang et al.) argues that agent memory should be distilled, typed knowledge rather than raw interaction text. We adapt that stance to a software repository's own git history under a constrained regime: deterministic, dependency-free, local-only, no embeddings. We present CommitDistill, an open-source Python prototype that mines a local git history into typed knowledge units (Facts, Skills, Patterns) using deterministic regex and surfaces them through a TF-IDF retriever with a calibrated silence threshold (theta = 2.5) that abstains on out-of-distribution queries. The artefact is a trust-instrumented memory substrate: deterministic, no external service, inspectable plain-JSON store, tunable abstention. A case study on five public repositories spanning Python, JavaScript, C, and Java (25,000 commits, 1,167 extracted units) reports useful-precision 0.525 at Cohen's kappa = 0.633 on 40 dual-annotated Python units. The decisive finding is budget-constrained retrieval: at a 256-character per-query budget, CommitDistill reaches 0.750 hit-rate on a 12-query benchmark against BM25's 0.333 and git log --grep's 0.083. On a four-arm paired LLM-as-judge evaluation (n=200 time-travel bug-fixes, two judges) covering control, CommitDistill, a body-budget-matched CD-Hybrid, and BM25, no condition produces a statistically detectable lift over control on the headline mean and CD-Hybrid is indistinguishable from BM25 head-to-head. Extraction over 10,000 commits completes in under 4 seconds on a laptop. Source, annotations, baselines, and a reproducibility script accompany this paper.", "published": "2026-05-18", "updated": "2026-05-18", "authors": [ "Divya Chukkapalli", "Thejesh Avula", "Aditya Aggarwal", "Harsimran Singh", "Amith Tallanki" ], "categories": [ "cs.SE", "cs.AI" ], "url": "https://arxiv.org/abs/2605.18284", "pdf_url": "https://arxiv.org/pdf/2605.18284", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.18421", "title": "EvoMemBench: Benchmarking Agent Memory from a Self-Evolving Perspective", "abstract": "Recent benchmarks for Large Language Model (LLM) agents mainly evaluate reasoning, planning, and execution. However, memory is also essential for agents, as it enables them to store, update, and retrieve information over time. This ability remains under-evaluated, largely because existing benchmarks do not provide a systematic way to assess memory mechanisms. In this paper, we study agent memory from a self-evolving perspective and introduce EvoMemBench, a unified benchmark organized along two axes: memory scope (in-episode vs. cross-episode) and memory content (knowledge-oriented vs. execution-oriented). We compare 15 representative memory methods with strong long-context baselines under a standardized protocol. Results show that current memory systems are still far from a general solution: long-context baselines remain highly competitive, memory helps most when the current context is insufficient or tasks are difficult, and no single memory form works consistently across all settings. Retrieval-based methods remain strong for knowledge-intensive settings, whereas procedural and long-term memory methods are more effective for execution-oriented tasks when their stored experience matches the task structure. We hope EvoMemBench facilitates future research on more effective memory systems for LLM-based agents. Our code is available at https://github.com/DSAIL-Memory/EvoMemBench.", "published": "2026-05-18", "updated": "2026-06-15", "authors": [ "Yuyao Wang", "Zhongjian Zhang", "Mo Chi", "Kaichi Yu", "Yuhan Li", "Miao Peng", "Bing Tong", "Chen Zhang", "Yan Zhou", "Jia Li" ], "categories": [ "cs.CL", "cs.AI", "cs.LG" ], "url": "https://arxiv.org/abs/2605.18421", "pdf_url": "https://arxiv.org/pdf/2605.18421", "matched_queries": [ "agent-memory", "long-term-memory" ] }, { "arxiv_id": "2605.18652", "title": "MementoGUI: Learning Agentic Multimodal Memory Control for Long-Horizon GUI Agents", "abstract": "Recent GUI agents have made substantial progress in visual grounding and action prediction, yet they remain brittle in long-horizon tasks that require maintaining task state across many interface transitions. Existing agents typically rely on raw history replay or text-only memory, which either overwhelms the model with redundant screenshots or discards localized visual evidence needed for future decisions. To address these limitations, we introduce \\textbf{MementoGUI}, a plug-in agentic memory framework that equips MLLM-based GUI agents with \\textbf{MementoCore}, a learned controller for online memory selection, compression, and retrieval. Rather than treating interaction history as a fixed context, MementoGUI formulates long-horizon GUI control as an online memory-control problem: working memory selectively preserves task-relevant interface events with textual summaries and ROI-level visual evidence, while episodic memory retrieves reusable past trajectories through learned relevance selection. MementoCore modularizes memory control into specialized operators for step processing, memory compression, episodic writing, and episodic selection, enabling plug-in memory augmentation without finetuning the GUI agent backbone. We further develop a scalable data curation pipeline that converts computer-use trajectories into memory-controller training data, introduce \\textbf{MementoGUI-Bench} for evaluating long-horizon decision-making in GUI agents, and design MLLM-based metrics for semantic action matching, task progress, and memory consistency. Experiments on GUI-Odyssey, MM-Mind2Web, and MementoGUI-Bench show that MementoGUI consistently improves GUI agents over no-history, history-replay, and text-only memory baselines, with larger MementoCore backbones further strengthening memory-augmented GUI control.", "published": "2026-05-18", "updated": "2026-05-18", "authors": [ "Ziyun Zeng", "Hang Hua", "Bocheng Zou", "Mu Cai", "Rogerio Feris", "Jiebo Luo" ], "categories": [ "cs.CV" ], "url": "https://arxiv.org/abs/2605.18652", "pdf_url": "https://arxiv.org/pdf/2605.18652", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.18930", "title": "OEP: Poisoning Self-Evolving LLM Agents via Locally Correct but Non-Transferable Experiences", "abstract": "Memory-augmented large language model (LLM) agents use iterative reflection and self-evolution to solve complex tasks, but these mechanisms introduce security risks. Existing agentic memory attacks require privileged access or explicit malicious content, making them detectable by advanced safety filters. This leaves a subtler attack surface underexplored: whether adversaries can induce agent to generate experiences that appear locally correct and semantically plausible yet induce harmful generalization during reflection. We find that reflective agents are vulnerable to such clean experiences, especially when paired with severe but plausible hypothetical consequences. Based on this observation, we introduce Obsessive Experience Poisoning (OEP), a low-privilege black-box attack requiring no direct control over the system prompt or memory database. OEP constructs adversarial clean edge-cases that combine locally correct solutions, non-transferable methods, and severe consequences, biasing reflection toward risk-averse rule formation. During memory consolidation, agents may over-trust self-generated reflections and distill localized experiences into high-priority but over-generalized rules, causing downstream failures. Evaluations across three domains show that OEP achieves ASR above 50\\% with GPT-4o agents, and outperforms existing attacks under LLM auditing defense.", "published": "2026-05-18", "updated": "2026-05-18", "authors": [ "Kaixiang Wang", "Jiong Lou", "Zhaojiacheng Zhou", "Jie Li" ], "categories": [ "cs.CR", "cs.AI", "cs.LG" ], "url": "https://arxiv.org/abs/2605.18930", "pdf_url": "https://arxiv.org/pdf/2605.18930", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.19952", "title": "Rethinking How to Remember: Beyond Atomic Facts in Lifelong LLM Agent Memory", "abstract": "To enable reliable long-term interaction, LLM agents require a memory system that can faithfully store, efficiently retrieve, and deeply reason over accumulated dialogue history. Most existing methods adopt an extracted fact based paradigm: handcrafted static prompts compress raw dialogues into atomic facts, which are then stored, matched, and injected into downstream reasoning. Nevertheless, such fact-centric designs inevitably discard fine-grained details in original dialogues and fail to support deep reasoning over scattered isolated facts. Moreover, static prompts cannot maintain consistent extraction granularity across diverse dialogue styles. To address these limitations, we propose TriMem, which maintains three coexisting representation granularities, including raw dialogue segments anchored by source identifiers for storage fidelity, extracted atomic facts for efficient memory retrieval, synthesized profiles that aggregate dispersed facts into holistic semantic understanding for deep reasoning. We further adopt TextGrad-based prompt optimization, which iteratively refines extraction and profiling prompts via response quality feedback, achieving lifelong evolution without any parameter updating. Extensive experiments on LoCoMo and PerLTQA across multiple LLM backbones demonstrate that TriMem consistently outperforms strong memory baselines. The code is available at https://TMLR-TriMem.github.io .", "published": "2026-05-19", "updated": "2026-05-19", "authors": [ "Jingwei Sun", "Jianing Zhu", "Jiangchao Yao", "Tongliang Liu", "Bo Han" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2605.19952", "pdf_url": "https://arxiv.org/pdf/2605.19952", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.20616", "title": "Auto-Dreamer: Learning Offline Memory Consolidation for Language Agents", "abstract": "Language agents increasingly operate over streams of related tasks, yet existing memory systems struggle to convert accumulated experience into reusable knowledge. Retrieval-augmented and structured memory methods record per-session observations effectively, but often couple acquisition and consolidation into a single online process, leaving the agent without a global view across sessions to discover recurring patterns, abstract shared procedures, or prune redundant entries. Inspired by complementary learning systems theory, we propose Auto-Dreamer, a learned offline consolidator for language-agent memory. Auto-Dreamer decouples fast per-session memory acquisition from slow cross-session consolidation. Given a selected working region of a typed memory bank, the consolidator treats the region as read-only evidence, performs bounded tool-use to inspect entries and provenance-linked source trajectories, and synthesizes a fresh compact replacement set that abstracts across sessions and supersedes the original region. We train Auto-Dreamer via GRPO, using end-to-end agent performance as the reward signal to learn how to consolidate memories acquired through fast online experience. Trained on ScienceWorld trajectories alone, Auto-Dreamer outperforms fixed, RL-trained, and prompted memory baselines on ScienceWorld by 7 points while using an active memory bank 12$\\times$ smaller than the strongest baseline, and continues to lead on held-out ALFWorld and WebArena without retraining -- using 6$\\times$ less memory than the strongest baseline on ALFWorld.", "published": "2026-05-20", "updated": "2026-05-20", "authors": [ "Chongrui Ye", "Yuxiang Liu", "Yu Wang", "Haofei Yu", "Yining Zhao", "Ge Liu", "Julian McAuley", "Jiaxuan You" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2605.20616", "pdf_url": "https://arxiv.org/pdf/2605.20616", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.20724", "title": "CALMem : Application-Layer Dual Memory for Conversational AI", "abstract": "Large language models (LLMs) operate within fixed context windows that fundamentally limit conversational continuity. When context fills, compaction discards history irreversibly; when sessions end, all memory resets to zero. Existing solutions-larger context windows, retrieval-augmented generation for knowledge bases, and memory-augmented architectures such as MemGPT-either require model modification, impose provider lock-in, or do not address the compaction continuity problem. We present CALMem (Conversational Application-Layer Memory), an application-layer dual memory architecture that gives LLM-based conversational assistants virtually unbounded effective context without any modification to the underlying model. CALMem combines two complementary memory subsystems: an episodic memory layer built on sliding-window vector embeddings of conversation history, and a semantic memory layer of agent-writable structured facts. A token-budget-adaptive injection mechanism, called the MOIM (Message of Injected Memory), automatically retrieves and injects relevant past context each turn, scaling injection depth inversely with context pressure. A key contribution is intra-session retrieval: compacted away turns from the current session remain searchable, closing a gap unaddressed by prior work. The system is implemented as a pure application layer in a production Rust codebase, is provider-agnostic, and degrades to original LLM behaviour with zero overhead when disabled. We describe the architecture, design decisions, and performance characteristics, and analyse the trade-offs that guided each implementation choice.", "published": "2026-05-20", "updated": "2026-05-20", "authors": [ "Rajendra Narayan Jena", "Rajan Padmanabhan", "Sankar Arumugam" ], "categories": [ "cs.IR" ], "url": "https://arxiv.org/abs/2605.20724", "pdf_url": "https://arxiv.org/pdf/2605.20724", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2605.20833", "title": "MemGym: a Long-Horizon Memory Environment for LLM Agents", "abstract": "Memory is a central capability for LLM agents operating across long-horizon tasks. Existing memory benchmarks predominantly evaluate retention of personalized information in multi-turn chat scenarios, overlooking the dynamic memory formation that occurs during extended agent execution. Consequently, the memory systems they produce transfer poorly to realistic agentic environments, such as coding and web navigation. We present MemGym, a benchmark for agentic memory that unifies existing agent gyms and in-house memory-grounded pipelines behind one memory-reasoning interface. MemGym spans five evaluation tracks grouped into four agentic regimes: tool-use dialogue (tau2-bench), multi-turn deep-research search (MEMGYM-DR), coding (SWE-Gym and MEMGYM-CODEQA), and computer use (WebArena-Infinity). MemGym reports memory-isolated scores that decouple memory performance from reasoning, retrieval, and tool-use ability, so memory strategies can be ranked without those confounders. Our synthetic pipelines for MEMGYM-CODEQA and MEMGYM-DR are length-controllable, ablation-verified at every stage, and tightly aligned with downstream scenarios. To make evaluation on coding environments academically tractable, we train MemRM, a lightweight reward model (Qwen3-1.7B fine-tuned with QLoRA) that scores compression quality as a fast scalar read in place of full Docker rollouts.", "published": "2026-05-20", "updated": "2026-05-20", "authors": [ "Wujiang Xu", "Yu Wang", "Kai Mei", "Kaiqu Liang", "Zhenting Wang", "Mingyu Jin", "Han Zhang", "Shi-Xiong Zhang", "Wenyue Hua", "Sambit Sahu", "Dimitris N. Metaxas" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2605.20833", "pdf_url": "https://arxiv.org/pdf/2605.20833", "matched_queries": [ "agent-memory", "memory-evaluation" ] }, { "arxiv_id": "2605.21463", "title": "Mem-$π$: Adaptive Memory through Learning When and What to Generate", "abstract": "We present Mem-$π$, a framework for adaptive memory in large language model (LLM) agents, where useful guidance is generated on demand rather than retrieved from external memory stores. Existing memory-augmented agents typically rely on similarity-based retrieval from episodic memory banks or skill libraries, returning static entries that often misalign with the current context. In contrast, Mem-$π$ uses a dedicated language or vision-language model with its own parameters, separate from the downstream agent, to generate context-specific guidance for complex tasks. Conditioned on the current agent context, the model jointly decides when to produce guidance and what guidance to produce. We train it with a decision-content decoupled reinforcement learning (RL) objective, enabling it to abstain when generation would not help and otherwise produce concise, useful guidance. Across diverse agentic benchmarks spanning web navigation, terminal-based tool use, and text-based embodied interaction, Mem-$π$ consistently outperforms retrieval-based and prior RL-optimized memory baselines, achieving over 30% relative improvement on web navigation tasks.", "published": "2026-05-20", "updated": "2026-05-20", "authors": [ "Xiaoqiang Wang", "Chao Wang", "Hadi Nekoei", "Christopher Pal", "Alexandre Lacoste", "Spandana Gella", "Bang Liu", "Perouz Taslakian" ], "categories": [ "cs.CL", "cs.AI" ], "url": "https://arxiv.org/abs/2605.21463", "pdf_url": "https://arxiv.org/pdf/2605.21463", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2605.22411", "title": "DeferMem: Query-Time Evidence Distillation via Reinforcement Learning for Long-Term Memory QA", "abstract": "Large language model (LLM) agents still struggle with long-term memory question answering, where answer-supporting evidence is often scattered across long conversational histories and buried in substantial irrelevant content. Existing memory systems typically process memory before future queries are known, then retrieve the resulting units based on similarity rather than their utility for answering the query. This workflow leaves downstream answerers to denoise retrieved candidates and reconstruct query-specific evidence. We present DeferMem, a long-term memory framework that decouples this problem into high-recall candidate retrieval and query-conditioned evidence distillation. DeferMem uses a lightweight segment-link structure to organize raw history and retrieve broad candidates at query time. It then applies a memory distiller trained with DistillPO, our reinforcement learning algorithm for distilling the high-recall but highly noisy candidates into a set of faithful, self-contained, and query-conditioned evidence. DistillPO formulates post-retrieval evidence distillation as a structured action comprising message selection and evidence rewriting. It optimizes this action with a decomposed-and-gated reward pipeline and structure-aligned advantage assignment, gating reward components from validity to quality checks while exposing task-level correctness feedback early and assigning each reward to its responsible output span. On LoCoMo and LongMemEval-S, DeferMem surpasses strong baselines in QA accuracy and memory-system efficiency, achieving the highest QA accuracy with the fastest runtime and zero commercial-API token cost for memory operations.", "published": "2026-05-21", "updated": "2026-05-21", "authors": [ "Jianing Yin", "Tan Tang" ], "categories": [ "cs.CL", "cs.AI", "cs.LG" ], "url": "https://arxiv.org/abs/2605.22411", "pdf_url": "https://arxiv.org/pdf/2605.22411", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2605.22721", "title": "Self-Evolving Multi-Agent Systems via Decentralized Memory", "abstract": "Self-evolving multi-agent systems (MAS) have emerged as a promising route to LLM agents that continually improve from experience, with persistent memory at their foundation. However, existing designs almost exclusively adopt a centralized repository shared across agents, incurring communication and coordination overhead, raising privacy concerns, and collapsing agent diversity. We propose DecentMem, a decentralized memory framework in which each agent maintains its own dual-pool memory -- an exploitation pool of consolidated past trajectories and an exploration pool of LLM-generated candidates for unseen contexts. The two pools are reweighted online based on stage-wise feedback from an LLM-as-a-judge. Theoretically, we prove that this design guarantees global reachability of the solution space and achieves $O(\\log T)$ cumulative regret, matching the stochastic bandit lower bound up to constants. In practice, across three MAS frameworks (AutoGen, DyLAN, AgentNet), three Qwen3 backbones (4B/8B/14B), two Gemma4 backbones (E2B/E4B) and five benchmarks spanning math, code, QA, and embodied tasks, DecentMem improves average accuracy by up to 23.8% over the strongest centralized memory baseline and by up to 52.5% over the no-memory baseline, while reducing token usage by up to 49%.", "published": "2026-05-21", "updated": "2026-05-21", "authors": [ "Guangya Hao", "Yunbo Long", "Zhuokai Zhao" ], "categories": [ "cs.MA" ], "url": "https://arxiv.org/abs/2605.22721", "pdf_url": "https://arxiv.org/pdf/2605.22721", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2605.23067", "title": "What Training Data Teaches RL Memory Agents: An Empirical Study of Curriculum Effects in Memory-Augmented QA", "abstract": "Reinforcement learning (RL) has emerged as a viable recipe for training LLM agents to reason over external memory banks in multi-session dialogue. Existing work trains exclusively on a single benchmark, leaving open how the composition of training data shapes the skills a memory agent acquires. We present a controlled empirical study that holds architecture, RL algorithm, and all hyperparameters fixed and varies only the training curriculum across three conditions: in-domain (LoCoMo), mixed-benchmark (LoCoMo + LongMemEval), and out-of-domain (LongMemEval only). Across two benchmarks and ten question types, curriculum composition acts as a fine-grained lever on specialization rather than a uniform scaling factor on performance. The mixed curriculum yields the strongest overall F1 on both evaluation sets. Training on a narrow out-of-domain set transfers a targeted skill - temporal reasoning - despite weak aggregate performance. Per-type differences substantially exceed aggregate differences, indicating that single-number benchmark comparisons systematically underreport curriculum effects. We further report two practical lessons from adapting GRPO to a single-GPU regime: cross-benchmark mixing requires filtering format-specific noise from memory banks to preserve training signal, and binary exact-match reward produces no learning signal at the small group sizes (G = 4) required on one GPU, motivating continuous reward functions in this regime.", "published": "2026-05-21", "updated": "2026-05-21", "authors": [ "Xinjie He", "Zhiyuan Lin", "Su Liu", "Jialun Wu", "Qiyang Xie", "Weikai Zhou", "Shuai Xiao" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2605.23067", "pdf_url": "https://arxiv.org/pdf/2605.23067", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.23723", "title": "MemAudit: Post-hoc Auditing of Poisoned Agent Memory via Causal Attribution and Structural Anomaly Detection", "abstract": "Large language model agents increasingly rely on persistent memory to store past interactions, retrieve relevant demonstrations, and improve long-horizon task execution. However, this memory mechanism also creates a practical security vulnerability: an adversarial user may inject malicious records into the agent's memory through ordinary interaction, and these records can later be retrieved to steer the agent's reasoning and actions. Existing defenses primarily focus on online intervention, such as prompt filtering or output blocking, but they do not address the post-hoc question of which stored memories are responsible after harmful behavior has already been observed. We propose \\textbf{MemAudit}, a post-hoc causal memory auditing framework for memory-augmented LLM agents. The framework combines two complementary signals: (1) a counterfactual memory influence score that measures each memory's causal contribution to harmful outputs, and (2) a memory consistency graph that identifies structurally anomalous memories within the broader memory store. We evaluate MemAudit against MINJA, a query-only memory injection attack in which malicious records are generated and stored through normal agent interactions rather than direct memory-bank modification. Across both QA and reasoning-agent settings, MemAudit substantially reduces attack success rates under realistic post-hoc auditing scenarios. The results show that QA attack success is reduced from $70\\%$ to $0\\%$, while RAP attack success drops from $83.3\\%$ to $0\\%$.", "published": "2026-05-22", "updated": "2026-05-22", "authors": [ "Zhewen Tan", "Yilun Yao", "Huiyan Jin", "Wenhan Yu", "Guoan Wang", "Mengyuan Fan", "liang lu", "Feng Liu", "Xiangzheng Zhang", "Duohe Ma", "Tong Yang", "Lin Sun" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2605.23723", "pdf_url": "https://arxiv.org/pdf/2605.23723", "matched_queries": [ "agent-memory", "long-term-memory" ] }, { "arxiv_id": "2605.24941", "title": "Memory-Induced Tool-Drift in LLM Agents", "abstract": "Modern LLM agents combine long-term memory for personalization with tool-calling interfaces for taking actions in the world -- a combination underpinning contemporary production systems. We study a previously unexamined failure of this combination: when personality-driven biases stored in memory (cost-consciousness, impatience, risk tolerance, etc.) silently affect tool calls in contexts where they are not applicable. We call this memory-induced tool-drift and operationalize it through MEMDRIFT, a benchmark of 105 scenarios spanning five bias dimensions and seven professional domains, generated through an automated adversarial pipeline. Across seven frontier models -- including those with extended reasoning -- biased memories raise deflection scores (a judge-scored measure of parameter deviation from unbiased baselines) by up to $+3.6$ points on a 1--5 scale. Tool-drift persists when memory management is handled by three production memory architectures. The phenomenon affects real-world tools: scanning 6{,}062 tools across 288 verified MCP servers, we flag 608 with susceptible parameters and confirm tool-drift on a validated subset. Mechanistically, biased memories act as implicit steering vectors, pushing activations along the same latent directions as explicit behavioral instructions. They also redistribute attention from task-relevant context toward memory entries with surface-level keyword overlap to the target parameter. Standard defenses -- prompt-based relevance instructions and memory filters -- reduce drift but do not eliminate it. As agents take increasingly consequential actions on a user's behalf, memory-induced tool-drift represents a systematic vulnerability that current safeguards do not address, motivating dedicated defenses at the intersection of memory management and tool-call generation.", "published": "2026-05-24", "updated": "2026-05-24", "authors": [ "Mahavir Dabas", "Jihyun Jeong", "Ming Jin", "Ruoxi Jia" ], "categories": [ "cs.CR", "cs.LG" ], "url": "https://arxiv.org/abs/2605.24941", "pdf_url": "https://arxiv.org/pdf/2605.24941", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2605.25002", "title": "MemMark: State-Evolution Attribution Watermarking for Agent Long-Term Memory Systems", "abstract": "Memory-backed agents need provenance that can survive leaked or migrated snapshots, where logs, visible outputs, and trusted metadata may be absent. We propose MemMark, a state-evolution attribution watermark that embeds an owner-controlled signal into latent memory-write decisions. At each internal LLM call, MemMark samples among admissible candidates using keyed, distribution-preserving selection, and records cryptographic commitments with signed session anchors and reveal evidence. This makes attribution depend on reproducible backend behavior rather than mutable provenance fields. Across A-Mem and Graphiti on LoCoMo, with three LLM backbones, MemMark preserves memory utility: Overall F1 retains 99.6% of the unwatermarked baseline, while BLEU-1 changes by +0.2%. It also provides usable carrier capacity, with 1.16, 1.14, and 1.26 bits of mean entropy for update-target, link-target, and semantic-realization decisions. In the snapshot-only R3 setting, MemMark recovers the full 40-bit payload from final snapshots, while wrong-key verification remains near chance. Under nine memory-lifecycle attacks, verification distinguishes tampering, evidence deletion, and partial payload recovery. These results show that robust snapshot-only attribution is feasible for long-term agent memory without surviving traces, trusted metadata, or utility-degrading.", "published": "2026-05-24", "updated": "2026-05-26", "authors": [ "Haobo Zhang", "Xutao Mao", "Guangyuan Dong", "Ziwei Li", "Xuanbo Su", "Kaijie Chen", "Jing Yang", "Zheng Lin" ], "categories": [ "cs.CR" ], "url": "https://arxiv.org/abs/2605.25002", "pdf_url": "https://arxiv.org/pdf/2605.25002", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.26154", "title": "MemMorph: Tool Hijacking in LLM Agents via Memory Poisoning", "abstract": "LLM-driven agents are capable of selecting external tools to complete users' tasks. However, attackers could compromise such process, steering agents toward inappropriate/wrong tools and enabling malicious actions. Most existing attacks primarily manipulate the tool metadata, which is easily detectable by auditing and may lose effectiveness as modern agents increasingly adopt memory modules to refine tool selection policies through accumulated experience. This paper proposes MemMorph, the first attack that bias tool selection by poisoning the agent's long-term memory. Rather than explicitly dictating the tool invocation decision, MemMorph injects a small number of crafted records that are disguised as technical facts, incident reports, and operational policies. These poisoned records reshape the agent's contextual perception and decision-making process, leading it to autonomously infer and select the tool preferred by the attacker. Experiments across 3 benchmarks, 10 agent backbones, and 3 memory-module implementations show that MemMorph achieves up to 85.9% attack success rate with only three injected records, outperforming the strongest baseline by up to 25% while retaining potency under 3 representative defenses. Our findings expose long-term memory as a critical and under-explored attack surface in tool-augmented agents, urging the development of memory-level integrity safeguards.", "published": "2026-05-24", "updated": "2026-05-24", "authors": [ "Xuanye Zhang", "Yongsen Zheng", "Zhuqin Xu", "Kaiyu Zhou", "Bowen Shen", "Haoran Ou", "Tianwei Zhang", "Kwok-Yan Lam" ], "categories": [ "cs.CR", "cs.AI" ], "url": "https://arxiv.org/abs/2605.26154", "pdf_url": "https://arxiv.org/pdf/2605.26154", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2605.25869", "title": "Mitigating Provenance-Role Collapse in Long-Term Agents via Typed Memory Representation", "abstract": "Long-term memory is essential for persistent LLM agents, yet prevailing architectures store historical interactions as unstructured, flat text. This unconstrained storage induces provenance-role collapse, a critical failure mode where agents suffer from source-monitoring errors. To resolve this cognitive vulnerability at the architectural level, we propose MemIR, a typed Memory Intermediate Representation that operationalizes source monitoring as a structural constraint. MemIR writes long-term memory into grounded atoms that separate raw evidence, retrieval cues, and truth-bearing claims, with factual authorization restricted to supported claim atoms. It then applies multi-route atomic projection and provenance-scoped utilization to transform heterogeneous retrieval hits into claim-centered candidate bundles and a normalized fact interface for answer generation. Experiments on LoCoMo and BEAM-100K demonstrate that MemIR consistently outperforms existing memory baselines, especially on tasks requiring source tracking, temporal grounding, and aggregation of fragmented evidence.", "published": "2026-05-25", "updated": "2026-05-25", "authors": [ "Zhengda Jin", "Bingbing Wang", "Jing Li", "Ruifeng Xu", "Min Zhang" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2605.25869", "pdf_url": "https://arxiv.org/pdf/2605.25869", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2605.26252", "title": "Is Agent Memory a Database? Rethinking Data Foundations for Long-Term AI Agent Memory", "abstract": "Long-running AI agents need persistent memory. Memory supports learning across sessions, reduces repeated context injection, and enables auditing of past decisions. Current agent memory systems and database paradigms treat memory as storage. They localize correctness at records, embeddings, or edges. Each supplies only some of the capabilities that long-term memory requires. The result is four recurring failure modes: unregulated growth, missing semantic revision, capacity-driven forgetting, and read-only retrieval. In our vision, long-term agent memory is a new data-management workload. Its correctness is a property of the state trajectory, not of individual records. We formalize this as Governed Evolving Memory (GEM). GEM replaces record-level database operations with four state-level operators: ingestion, revision, forgetting, and retrieval. Six correctness conditions govern how the state evolves. Three structural observations establish that no record-level system can satisfy these conditions, regardless of the storage model. We realize the abstraction in MemState, a prototype on a property-graph backend. MemState validates feasibility and exposes the gap to a native engine. We outline three research directions that define memory-centric data management as a workload.", "published": "2026-05-25", "updated": "2026-05-25", "authors": [ "Abdelghny Orogat", "Essam Mansour" ], "categories": [ "cs.AI", "cs.DB" ], "url": "https://arxiv.org/abs/2605.26252", "pdf_url": "https://arxiv.org/pdf/2605.26252", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.26256", "title": "Personalizing Embodied Multimodal Large Language Model Agents over Long-term User Interactions", "abstract": "Multimodal large language model (MLLM)-based embodied agents have shown strong potential for solving complex tasks in physical environments. However, personalized assistance requires more than following generic instruction or recognizing object categories. In real-world scenarios, the intended target is often specified only implicitly through prior interactions, requiring agents to leverage personalized context accumulated over time. In this work, we propose POLAR, a multiomodal memory-augmented framework for personalized embodied agents over long-term user interactions. POLAR organizes prior interactions into a multimodal knowledge graph that captures semantic memory for personalized context and visual concepts, and episodic memory for embodied experiences such as agent trajectories. To execute embodied tasks, POLAR retrieves relevant memories to interpret the current request and guide task execution. We evaluate POLAR across multiple MLLM backbones and diverse evaluation scenarios to study the role of memory in long-term personalization. Results show that the proposed memory mechanism consistently improves performance by enabling more effective use of information accumulated over prior interactions. The gains are especially pronounced when the agents are required to reason across multiple interactions, perform multi-hop inference, or tracking updates in user-specific context over time.", "published": "2026-05-25", "updated": "2026-05-25", "authors": [ "Jeongeun Lee", "Chanyoung Park", "Dongha Lee" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2605.26256", "pdf_url": "https://arxiv.org/pdf/2605.26256", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2605.27366", "title": "MUSE-Autoskill: Self-Evolving Agents via Skill Creation, Memory, Management, and Evaluation", "abstract": "Large language model (LLM) agents rely on reusable skills to solve complex tasks, but existing skill creation approaches often treat skills as isolated, static artifacts, limiting reusability, reliability, and long-term improvement. We propose MUSE-Autoskill Agent (Memory-Utilizing Skill Evolution), a skill-centric agent framework that creates, reuses, and refines skills under a unified lifecycle: creation, memory, management, evaluation, and refinement. MUSE creates skills on demand, stores them across tasks, retrieves them through a skill catalog, and accumulates per-skill experience for later reuse and adaptation. Across the main reported settings on SkillsBench and SkillLearnBench, MUSE-Autoskill outperforms Hermes, Codex, and Claude Code. On SkillsBench, its self-created skills surpass human-authored skills on the successfully covered subset (85.24% vs. 81.17%), showing that lifecycle-managed skills can distill agent experience into highly effective reusable assets; MUSE-created skills also transfer to Hermes more effectively than Codex- or Claude-created skills, reaching 51.90% accuracy under transfer. These results highlight the importance of treating skills as long-lived, experience-aware, and testable assets.", "published": "2026-05-26", "updated": "2026-07-03", "authors": [ "Huawei Lin", "Peng Li", "Jie Song", "Fuxin Jiang", "Tieying Zhang" ], "categories": [ "cs.AI", "cs.CL", "cs.LG", "cs.MA" ], "url": "https://arxiv.org/abs/2605.27366", "pdf_url": "https://arxiv.org/pdf/2605.27366", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.27760", "title": "SkillGrad: Optimizing Agent Skills Like Gradient Descent", "abstract": "Agent skills provide a lightweight way to adapt LLM agents to specialized domains by storing reusable procedural knowledge in structured files. However, whether downloaded from third parties or self-generated, these skills are often unreliable, incomplete, or outdated. Existing skill-evolution methods often address these deficiencies through heuristic reflections without an explicit optimization formulation. In this paper, we propose SkillGrad, a gradient-descent-inspired framework for optimizing agent skills. SkillGrad treats the skill package as a structured parameter to optimize in a gradient descent fashion: task executions provide trajectory-level loss evidence, automatic diagnoses then provide text-based gradients that indicate the correction directions. To stabilize optimization across iterations, a momentum agent accumulates recurring diagnostic patterns into a persistent memory overlay. Finally, an LLM-based patcher executes the parameter update by applying layer-aware edits to the skill package. Evaluated on SpreadsheetBench Verified and WikiTableQuestions, SkillGrad consistently outperforms training-based skill evolution baselines across two backbone LLMs, improving over the strongest training-based baseline by $6.7$ percentage points on average. Ablations further show that momentum and contrastive diagnosis both contribute to the final skill quality.", "published": "2026-05-26", "updated": "2026-05-26", "authors": [ "Hanyu Wang", "Yifan Lan", "Bochuan Cao", "Lu Lin", "Jinghui Chen" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2605.27760", "pdf_url": "https://arxiv.org/pdf/2605.27760", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2605.27762", "title": "PEAM: Parametric Embodied Agent Memory through Contrastive Internalization of Experience in Minecraft", "abstract": "We present PEAM, a Parametric Embodied Agent Memory framework in Minecraft that transforms agent memory from inference-time retrieval into parameter-resident skills internalized through experience. PEAM pairs a slow deliberative LLM for open-ended reasoning with a fast parametric module for reflexive execution of consolidated skills. The fast module is a multimodal Mixture-of-Experts LoRA architecture with per-category physically isolated adapters, enabling parameter-level continual learning without catastrophic forgetting. We treat failure as a first-class training signal: failure--correction trajectory pairs are internalized through a joint behavioral-cloning and contrastive objective, so the agent learns not only what succeeds but also how corrected actions differ from failed ones. To govern consolidation, PEAM introduces a parameterization-worthiness score for deciding which experience should be internalized, and a scale-free self-triggered consolidation mechanism for deciding when to internalize without task-specific hand-tuned thresholds, making the agent self-evolving as the trigger transfers across task distributions without re-tuning. Experiments in Minecraft show that PEAM improves long-horizon task performance, mitigates forgetting on previously consolidated skills, and improves parametric-versus-retrieval efficiency over retrieval-based embodied agents and parametric memory variants.", "published": "2026-05-26", "updated": "2026-06-01", "authors": [ "Yuchen Guo", "Junli Gong", "Weicheng Wang", "Hongmin Cai", "Yiu-ming Cheung", "Weifeng Su" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2605.27762", "pdf_url": "https://arxiv.org/pdf/2605.27762", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.27825", "title": "MRMMIA: Membership Inference Attacks on Memory in Chat Agents", "abstract": "Membership inference attacks (MIAs) test whether a target data record belongs to a system's private data, and have become a standard tool to measure privacy leakage in machine learning systems. Prior work has primarily focused on training corpora or retrieval databases. However, MIAs against agent memory have received less attention, even though such memory can contain sensitive user-agent interactions, retrieved facts, and user preferences. Therefore, in this work, we focus on chat agent memory MIAs, where an adversary infers whether a candidate memory unit belongs to the chat agent's memory store. We propose Multi-Recall Memory MIA (MRMMIA), a unified attack that utilizes multiple recall probes to the agent to extract the membership signal across black-box, gray-box, and white-box settings. Our experiments demonstrate that MRMMIA consistently outperforms baselines. Our results expose the privacy risk in agents and provide an initial evaluation framework for membership leakage in chat-agent memory systems.", "published": "2026-05-27", "updated": "2026-05-27", "authors": [ "Kai Chen", "Yan Pang", "Tianhao Wang" ], "categories": [ "cs.CR", "cs.LG" ], "url": "https://arxiv.org/abs/2605.27825", "pdf_url": "https://arxiv.org/pdf/2605.27825", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.28046", "title": "MemCog: From Memory-as-Tool to Memory-as-Cognition in Conversational Agents", "abstract": "Existing agent memory systems universally follow what we term a Memory-as-Tool paradigm where a single query triggers one-shot retrieval of flat passage lists, suffering from passive invocation, reasoning-retrieval decoupling, and structural mismatch between retrieved fragments and the agent's navigational needs. We propose MemCog, a Memory-as-Cognition system that makes memory access an integral part of the reasoning process. MemCog organizes user knowledge as Navigable Memory Store with associative link graphs, exposes Cross-Dimensional Navigation Interface for multi-step reasoning-driven traversal, and employs Proactive Reasoning Protocol that drives agents to spontaneously initiate memory exploration from conversational context. We additionally construct ProactiveMemBench, the first benchmark for evaluating proactive memory triggering. Experiments show that MemCog achieves state-of-the-art on passive QA benchmarks (92.98 on LoCoMo, 95.8 on LongMemEval) while substantially outperforming baselines on ProactiveMemBench, demonstrating the advantage of Memory-as-Cognition.", "published": "2026-05-27", "updated": "2026-05-27", "authors": [ "Zihan Li", "Xingyu Fan", "Feifei Li", "Wenhui Que" ], "categories": [ "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2605.28046", "pdf_url": "https://arxiv.org/pdf/2605.28046", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.29341", "title": "WorldMemArena: Evaluating Multimodal Agent Memory Through Action-World Interaction", "abstract": "Multimodal large language models are increasingly deployed as long-horizon agents, where memory must do more than recall: it must track an evolving world, revise what has gone stale, and surface the right evidence at decision time. Existing benchmarks measure recall over static dialogue, collapse memory into a single end-of-task accuracy, and reduce visual observations to captions, leaving us unable to localize failures to writing, maintenance, retrieval, or use. The rise of agent harnesses that author their own memory sharpens this gap, since we have no principled way to compare hand-designed pipelines with self-managing alternatives. To close these gaps, we formulate multimodal agent memory as an Action-World Interaction Loop with an observable four-stage lifecycle, and instantiate it in WorldMemArena: 400 multi-session multimodal tasks spanning Lifelong Evolution (evolving personal and task states) and Agentic Execution (memory from real observations, actions, and feedback), annotated with gold memory points, updates, distractors, and evidence chains for stage-level diagnosis. This enables the first head-to-head comparison of long-context, manually designed (RAG and external memory systems), and harness-based memory agents. Results show that: (1) better memory writing and storage do not guarantee better performance; (2) multimodal memory still struggles to fully use visual evidence; (3) systems are unstable across domains and degrade on realistic agentic trajectories; and (4) harness memory is more flexible but remains costly and less reliable.", "published": "2026-05-28", "updated": "2026-06-01", "authors": [ "Chengzhi Liu", "Yuzhe Yang", "Sophia Xiao Pu", "Yepeng Liu", "Lin Long", "Yichen Guo", "Nuo Chen", "Zhaotian Weng", "Elena Kochkina", "Simerjot Kaur", "Charese Smiley", "Xiaomo Liu", "James Zou", "Sheng Liu", "Yuheng Bu", "Songyou Peng", "Xin Eric Wang" ], "categories": [ "cs.CV", "cs.CL" ], "url": "https://arxiv.org/abs/2605.29341", "pdf_url": "https://arxiv.org/pdf/2605.29341", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.29630", "title": "Entity-Collision: A Stratified Protocol for Attributing Retrieval Lift in Agent Memory", "abstract": "End-to-end agent-memory benchmarks report a single hit@k per retriever, confounding lexical leakage (uncontrolled query/gold/distractor entity overlap) with tag-mixing (preferences, services, tools averaged together). We propose entity-collision, a system-agnostic protocol that pins the BM25 floor by construction -- every distractor shares the answer's entity tokens -- and stratifies queries by discriminator tag, so any lift over BM25 is attributable to the embedder. Applied to an open-source agent-memory testbed across 5 tags x 3 embedders x 5 collision degrees with paired-bootstrap 95% CIs, the protocol reveals a two-axis pattern: a 256-d hash trigram helps only on closed-vocabulary lexical tags at deep collision; MiniLM-384 dominates both axes; and a 2.7x-parameter BGE-large does not uniformly improve on MiniLM -- it wins on intent-style queries but loses on lexical ones. Encoder capacity alone is not the binding constraint. The synthetic intent-tag null replicates on LongMemEval (n=500) as a single-session-preference recall cliff. Adaptive vector-weight routing on LoCoMo is a measured null: 11.7pp of oracle headroom exists, but no signal we tested recovers it. All 26 result tables and 37 reproduce scripts are version-controlled and verified by a public registry; the protocol is exercised on a deterministically governed memory testbed (event-sourced decision log, DAG-state-machine schema lifecycle) so every reported CI is reproducible byte-for-byte from the ingest stream.", "published": "2026-05-28", "updated": "2026-05-28", "authors": [ "Youwang Deng" ], "categories": [ "cs.CL", "cs.AI", "cs.IR" ], "url": "https://arxiv.org/abs/2605.29630", "pdf_url": "https://arxiv.org/pdf/2605.29630", "matched_queries": [ "agent-memory", "memory-evaluation" ] }, { "arxiv_id": "2605.29640", "title": "VikingMem: A Memory Base Management System for Stateful LLM-based Applications", "abstract": "Large Language Models have revolutionized interactive applications; however, their finite context windows pose a critical data management challenge for maintaining stateful, long-term interactions. Existing memory approaches often rely on simplistic extraction methods that lead to incomplete memories or use rigid, single-purpose memory extraction prompts tailored to a single use case, such as chatbots. Consequently, they lack generalizability and perform poorly across diverse downstream tasks. To bridge this gap, we introduce the Memory Base, a novel data management paradigm for managing the persistent state of long-term interactions. It is characterized by three core principles: selective extraction of high-value memories from raw information streams; inherent statefulness and evolution, where memory content is progressively summarized, corrected, and temporally weighted to prioritize recent interactions; and a generalizable abstraction paradigm designed for robust transferability across diverse applications, including education, recommendation, and agent memory. Building on this foundation, we present VikingMem, an end-to-end Memory Base Management System implemented on the VikingDB vector engine. VikingMem materializes this paradigm through interconnected event and entity abstractions. It features event-centric memory extraction to selectively handle complex information streams, while entities are dynamically updated by events to achieve stateful evolution. Using temporal compression via a topic-wise timeline and time-weighted recall, the system progressively produces high-level summary memories, prioritizes recent items, and compresses and fades older ones. Extensive evaluations on long-term memory benchmarks demonstrate that VikingMem outperformes baselines by up to 30% in memory retrieval effectiveness while maintaining the low latency essential for interactive applications.", "published": "2026-05-28", "updated": "2026-06-12", "authors": [ "Jiajie Fu", "Junwen Chen", "Mengzhao Wang", "Aoxiang He", "Maojia Sheng", "Xiangyu Ke", "Yifan Zhu", "Yunjun Gao" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2605.29640", "pdf_url": "https://arxiv.org/pdf/2605.29640", "matched_queries": [ "agent-memory", "memory-evaluation" ] }, { "arxiv_id": "2605.29960", "title": "Hijacking Agent Memory: Stealthy Trojan Attacks Through Conversational Interaction", "abstract": "Large language model (LLM) agents increasingly leverage long term memory to support persistent and autonomous task execution. However, this capability also introduces a new attack surface: memory poisoning, where adversaries can inject malicious information to influence future behavior. Existing memory poisoning attacks often assume that injected content can be stored directly in memory, overlooking the selective extraction and rewriting stages in modern memory pipelines. This makes prior methods ineffective under realistic settings. In this paper, we propose MemPoison, a novel memory poisoning attack that bypasses selective memory mechanisms in LLM agents, where an attacker can inject triggerable backdoors into the agent's long-term memory through dialogue interactions, thereby misleading its subsequent responses. MemPoison introduces three key components: (i) a semantic relational bridge that binds the trigger and payload into a coherent statement to ensure they are extracted into memory together; (ii) entity masquerading that optimizes triggers to mimic named entities, resisting rewriting; and (iii) joint embedding optimization that shapes trigger-injected texts into a tight cluster in the embedding space while maintaining isolation from benign embeddings for stealth. Evaluations across different agent domains and memory mechanisms show MemPoison achieves attack success rates up to 0.95, outperforming existing baselines. Mechanistic analysis indicates that the attack exploits embedding-space anisotropy and shifts attention patterns, highlighting core vulnerabilities in selective memory systems. We evaluate multiple defense strategies and demonstrate their fundamental limitations in mitigating the attack.", "published": "2026-05-28", "updated": "2026-05-28", "authors": [ "Hongtao Wang", "Se Yang", "Yu Chen", "Puzhuo Liu" ], "categories": [ "cs.CR", "cs.AI" ], "url": "https://arxiv.org/abs/2605.29960", "pdf_url": "https://arxiv.org/pdf/2605.29960", "matched_queries": [ "agent-memory", "long-term-memory" ] }, { "arxiv_id": "2605.30058", "title": "HEART-Bench: Do LLM Agents Exhibit Human-like Psychology?", "abstract": "While LLM agents have demonstrated remarkable task-oriented abilities such as planning, reasoning, and action, few works have treated them as complete human personalities where emotional dimensions hold equal importance. In this paper, we introduce a novel benchmark to systematically assess whether LLM agents can simulate coherent, human-like psychology. Specifically, our benchmark constructs 11 diverse human characters grounded in orthogonal Big Five personality traits, with each profile deeply integrated with 1,000 structured autobiographical-style episodic memories distributed across theory-grounded developmental life stages. To rigorously evaluate the psychological manifestations of LLMs, we designed a curated suite of 64 decision-making scenarios, guided by the DIAMONDS taxonomy, a psychological framework that characterizes situations along eight dimensions: Duty, Intellect, Adversity, Mating, pOsitivity, Negativity, Deception, and Sociality. By subjecting agents to varying scenarios, the benchmark evaluates whether they can consolidate their innate personality traits and autobiographical memories to make behavioral decisions that are consistent with their specific psychological profiles. After systematic human validation and filtering, we obtained a benchmark consisting of 673 multiple-choice questions (MCQs). We believe this benchmark provides a principled and scalable testbed for studying human-like emotions, personality consistency, and value-consistent behavioural decision-making in LLM-based agents.", "published": "2026-05-28", "updated": "2026-05-28", "authors": [ "Weihan Peng", "Chenxu Zhang", "Qianao Wang", "Yuling Shi", "Heng Lian", "Qihong Mao", "Jiahao Pang", "Chunliang Feng", "Bowen Li", "Xiaodong Gu" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2605.30058", "pdf_url": "https://arxiv.org/pdf/2605.30058", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2605.30621", "title": "Harness Updating Is Not Harness Benefit: Disentangling Evolution Capabilities in Self-Evolving LLM Agents", "abstract": "LLM agents are increasingly deployed as systems built around editable external harnesses, including prompts, skills, memories and tools, that shape task execution without changing model parameters. Harness self-evolution adapts such agents by updating these harnesses from execution evidence. Yet it remains unclear whether a model's base capability in task-solving predicts its capabilities in harness self-evolution: which models produce useful harness updates, and which actually benefit from them? We analyze two harness self-evolution capabilities: (i) harness-updating, the capability to produce useful persistent harness updates from execution evidence; (ii) harness-benefit, the capability to benefit from updated harnesses during task solving. Our analysis reveals two findings. First, harness-updating is flat in base capability: models from different capability tiers produce harness updates that lead to surprisingly similar gains; even Qwen3.5-9B's updates yield gains comparable to those of Claude Opus~4.6. Second, harness-benefit is non-monotonic in base capability: weak-tier models benefit little from updated harnesses, mid-tier models benefit most, and strong-tier models benefit less than mid-tier. We trace low gains at the weak tier to two failure modes: weak-tier models may fail to activate relevant harness artifacts, or activate them but fail to follow them faithfully. These findings suggest investing capability budget in the task-solving agent rather than the evolver, and targeting harness invocation and long-horizon instruction following in agent training. Our source code is publicly available at https://github.com/A-EVO-Lab/a-evolve/tree/release/harness-evolution.", "published": "2026-05-28", "updated": "2026-05-28", "authors": [ "Minhua Lin", "Juncheng Wu", "Zijun Wang", "Zhan Shi", "Yisi Sang", "Bing He", "Zewen Liu", "Tianxin Wei", "Zongyu Wu", "Zhiwei Zhang", "Dakuo Wang", "Xiang Zhang", "Benoit Dumoulin", "Cihang Xie", "Yuyin Zhou", "Suhang Wang", "Hanqing Lu" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2605.30621", "pdf_url": "https://arxiv.org/pdf/2605.30621", "matched_queries": [ "procedural-memory" ] }, { "arxiv_id": "2607.00017", "title": "Learning User-Aware Recall: Personalized Retrieval in Long-Term Conversational Memory", "abstract": "Long-term conversational agents are expected to remember past interactions, but memory is useful only when the right evidence is recalled for the right user. Existing memory-augmented LLM agents have made progress in building compact memory banks, yet retrieval is still often driven by query-centered similarity or fixed ranking rules, leaving user-conditioned relevance underexplored. To address this gap, we propose Profile-guided Personalized Retrieval Optimization (PPRO), a retrieval-centric framework that makes memory retrieval both user-aware and optimizable. PPRO builds episodic and semantic memory banks from dialogue histories and derives a user profile from accumulated memories. The profile serves as an explicit personalized prior in memory ranking, allowing retrieval to account for stable user attributes, preferences, and relationships. PPRO further trains a query rewriter with Group Relative Policy Optimization, using both evidence retrieval quality and downstream answer quality as feedback while keeping the memory banks and answer model fixed. Experiments on LoCoMo and LongMemEval-S show consistent gains over training-free memory systems and training-based baselines. Ablation studies further show that both profile-guided ranking and retrieval-oriented rewriting contribute substantially to performance, highlighting retrieval optimization as a key factor in personalized long-term memory use.", "published": "2026-05-28", "updated": "2026-07-02", "authors": [ "ZhiShu Jiang", "Haibo Liu", "Xin Shen", "Guanqiang QI", "Chenxi Miao", "Weikang Li", "Liwei Qian", "Xin Pei", "Jizhou Huang" ], "categories": [ "cs.IR", "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2607.00017", "pdf_url": "https://arxiv.org/pdf/2607.00017", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2605.30690", "title": "ElasticMem: Latent Memory as a Learnable Resource for LLM Agents", "abstract": "Long-term memory is essential for LLM agents to reason coherently across extended interactions, personalize responses, and reuse past experience. However, existing memory-augmented methods typically treat memory as a fixed resource: text-space approaches concatenate retrieved memories into the context window, causing substantial token overhead and sensitivity to noisy evidence, while latent-space approaches reduce textual cost but still rely on rigid retrieval or fixed-capacity memory interfaces. This creates a mismatch between query-dependent memory utility and fixed memory allocation. We propose ElasticMem, a memory-augmented LLM framework that learns to use memory as an elastic latent resource. ElasticMem builds an offline latent memory bank with retrieval keys and content caches, retrieves memories adaptively from the reasoner's hidden state, assigns each retrieved memory a variable latent budget through a learned policy, and injects selected latent states as soft memory tokens for generation. The full memory-use process is optimized with downstream task rewards through group-relative policy optimization. We evaluate ElasticMem on MemorySuite, covering memory-intensive QA and embodied agent control. Across Qwen2.5-3B-Instruct and Qwen2.5-7B-Instruct backbones, ElasticMem improves weighted average QA accuracy by 26.2% and 24.6%, and improves ALFWorld success rate by 66.3% and 27.2%, respectively, over the strongest baselines, while achieving the lowest ALFWorld token cost. Ablations and qualitative analyses further show that adaptive retrieval and elastic budget allocation help ElasticMem prioritize useful evidence and transferable plans beyond rigid cosine similarity. Our code for ElasticMem will be released at https://github.com/ulab-uiuc/ElasticMem.", "published": "2026-05-29", "updated": "2026-05-29", "authors": [ "Tao Feng", "Chongrui Ye", "Tianyang Luo", "Jingjun Xu", "Xueqiang Xu", "Haozhen Zhang", "Ge Liu", "Jiaxuan You" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2605.30690", "pdf_url": "https://arxiv.org/pdf/2605.30690", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2605.30711", "title": "SAGE: A Novelty Gate for Efficient Memory Evolution in Agentic LLMs", "abstract": "Agentic LLMs must continuously decide whether newly extracted facts should be added, merged with existing memories, or ignored, yet prior work has focused more on retrieval and storage than on principled write-side control. We frame memory evolution as a novelty-detection problem and propose SAGE, a Spherical Adaptive Gate for memory Evolution that scores candidate facts with a von Mises-Fisher-based density estimator over memory embeddings and routes them with an adaptive threshold that tracks memory-store geometry. SAGE resolves clearly novel facts as ADD, clearly redundant facts as NOOP, and sends only uncertain cases to an LLM merge step, reducing expensive write-time reasoning. On LoCoMo, SAGE achieves the best average token-F1 against Mem0 on all seven open-weight backbone comparisons, while on GPT-4o-mini it reduces add-phase API cost by 3.4$\\times$ and add-phase latency by 2.5$\\times$ with only a small average judge-score gap. As a drop-in binary gate for A-Mem, SAGE skips roughly 16-18% of LLM calls across five models with minimal quality change on open-weight backbones. These results suggest that novelty-aware write control is a practical lever for improving both memory quality and system efficiency in long-term agentic memory. The source code for our approach is accessible at https://github.com/swang1024/SAGE.", "published": "2026-05-29", "updated": "2026-06-18", "authors": [ "Sijia Wang", "Dhanajit Brahma", "Ricardo Henao" ], "categories": [ "cs.CL", "cs.AI", "cs.LG", "stat.ML" ], "url": "https://arxiv.org/abs/2605.30711", "pdf_url": "https://arxiv.org/pdf/2605.30711", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.30858", "title": "ForecastCompass: Guiding Agentic Forecasting with Adaptive Factor Memory", "abstract": "Agentic forecasting is important for decision-making in dynamic environments, but it remains challenging because agents must reason from incomplete, time-limited evidence and produce calibrated probabilities before outcomes are resolved. Memory provides a natural mechanism for transferring experience from resolved forecasts to future prediction tasks. However, existing agent-memory methods are not tailored to forecasting, as they typically store past interactions, reflections, or factual associations without explicitly representing reusable predictive factors or calibration knowledge. We propose ForecastCompass (FoCo), an adaptive factor-based memory framework for agentic forecasting. FoCo organizes forecasting experience with a hierarchical forecasting-task taxonomy, enabling retrieval task-relevant forecasting knowledge. It maintains two complementary memory components: factor memory, which captures reusable predictive dimensions, and reasoning memory, which encodes probability updating, uncertainty handling, and calibration principles. Using retrospective analyses as learning signals, FoCo iteratively revises memory through a verbalized memory-revision procedure, enabling the agent to accumulate transferable forecasting knowledge over time. Experiments on Prophet Arena and FutureX with GPT-5-mini and Gemini-2.5-Flash show that FoCo improves both probabilistic accuracy and calibration.", "published": "2026-05-29", "updated": "2026-05-29", "authors": [ "Yurui Chang", "Yongkang Du", "Yuanpu Cao", "Jinghui Chen", "Lu Lin" ], "categories": [ "cs.LG" ], "url": "https://arxiv.org/abs/2605.30858", "pdf_url": "https://arxiv.org/pdf/2605.30858", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2605.31075", "title": "Task-Focused Memorization for Multimodal Agents", "abstract": "Long-term memory is essential for multimodal agents to build coherent experience, accumulate world knowledge, and achieve continual learning. However, constructing effective memory goes beyond memory module design and basic requirements such as accuracy and fidelity; the key challenge lies in determining what to memorize. Multimodal agents, such as embodied agents, continuously perceive, reason, and act in real or virtual environments, receiving an unbounded stream of multimodal observations. From this combinatorial explosion of information, an agent must selectively retain content that is relevant to its role in the environment and valuable for future tasks. To bridge this gap, we frame memory generation as a learnable memorization policy and introduce TaskMem (Task-focused Memorization Policy Learning), a reinforcement-learning-based framework that enables the policy to dynamically adjust its focus to the demands of real tasks encountered in the environment. TaskMem adopts a two-phase training paradigm: Phase One learns how to memorize by optimizing memory quality under fundamental fidelity requirements; Phase Two occurs after deployment, where the agent learns what to memorize by tuning an adapter on its base MLLM, using recent environment tasks to define a reward model that guides the memorization policy toward task-relevant content. To evaluate our approach, we reformulate VideoMME, EgoLife, and EgoTempo into streaming benchmarks that simulate a realistic setting in which an agent processes streaming observations and handles tasks arriving online. To isolate memory assessment, the questions must be answered using only the agent's memory, without access to raw video. Built on Qwen3-VL-30B-A3B, TaskMem improves VQA accuracy by 6.3%, 7.0%, and 5.3% on these benchmarks, respectively.", "published": "2026-05-29", "updated": "2026-05-29", "authors": [ "Tao Zou", "Yichen He", "Tian Qiu", "Yuan Lin", "Hang Li" ], "categories": [ "cs.CV" ], "url": "https://arxiv.org/abs/2605.31075", "pdf_url": "https://arxiv.org/pdf/2605.31075", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.00619", "title": "MemPro: Agentic Memory Systems as Evolvable Programs", "abstract": "Long-horizon autonomous agents require memory systems to retain historical information, track evolving states, and reuse relevant knowledge beyond finite context windows. Existing agentic memory systems typically follow a memory construction-retrieval (MCR) pipeline, but often adapt mainly the memory bank while keeping the surrounding pipeline fixed after deployment. This fixed-pipeline design struggles to handle heterogeneous task-specific failure modes and can become misaligned with memory banks that evolve in scale and structure over time. To address these limitations, we propose MemPro, a system-level evolution framework that treats the entire MCR pipeline as an evolvable program rather than adapting only the memory bank or prompt text. MemPro maintains a version tree of runnable memory-system implementations, where an Evolving Agent iteratively selects promising versions, diagnoses recurring failures, and creates improved child versions through failure-mode-guided edit-debug refinement. Experiments on LongMemEval, LoCoMo, HotpotQA, and NarrativeQA show that MemPro consistently outperforms strong static and prompt-level evolving baselines within a few iterations, continues to improve with evolution, and achieves a favorable performance-cost trade-off. Code is available at https://github.com/wanghai673/MemPro.", "published": "2026-05-30", "updated": "2026-05-30", "authors": [ "Qingshan Liu", "Guoqing Wang", "Wen Wu", "Jingqi Huang", "Xinqi Tao", "Dejia Song", "Jie Zhou", "Liang He" ], "categories": [ "cs.CL", "cs.AI" ], "url": "https://arxiv.org/abs/2606.00619", "pdf_url": "https://arxiv.org/pdf/2606.00619", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.00734", "title": "EMA: Approximate Nearest Neighbor Search with General Attribute Filtering and Dynamic Updates", "abstract": "Filtering Approximate Nearest Neighbor (FANN) search is a critical and emerging task for strengthening the query capability of vector databases, supporting applications such as recommendation systems, retrieval-augmented generation (RAG), and agent memory. However, most existing methods are limited to range or label filtering, often incurring unacceptable index construction time and memory overhead. Predicate-agnostic approaches further struggle to handle a wide range of predicate selectivities effectively. In this paper, we propose EMA, a filtering ANN algorithm that supports multi-predicate queries over mixed numerical and categorical attributes, and efficient dynamic updates. EMA introduces Markers as compact summaries attached to graph edges, providing conservative predicate- and geometric-aware guidance with zero false negatives at the Marker level. During query processing, EMA performs Marker-augmented joint search with a bounded edge recovery mechanism, enabling efficient filtering while preserving graph navigability. Extensive experiments demonstrate that EMA achieves 1.68x--12.25x speedup over state-of-the-art general filtering ANN methods across diverse workloads.", "published": "2026-05-30", "updated": "2026-05-30", "authors": [ "Mocheng Li", "Baotong Lu", "James Cheng", "Chenhao Ma" ], "categories": [ "cs.DB" ], "url": "https://arxiv.org/abs/2606.00734", "pdf_url": "https://arxiv.org/pdf/2606.00734", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.00756", "title": "CoMIC: Collaborative Memory and Insights Circulation for Long-Horizon LLM Agents in Cloud-Edge Systems", "abstract": "Deploying lightweight Large Language Model (LLM) agents on edge servers can reduce latency and move agentic services closer to users, but resource-constrained edge models often struggle with long-horizon tasks that require persistent memory, subgoal tracking, and reflection. Fine-tuning edge models after deployment is costly and difficult to scale across heterogeneous nodes, while purely local memory leaves agents with isolated experience and growing prompt context. We propose \\textsc{CoMIC}, a parameter-update-free cloud-edge framework for Collaborative Memory and Insights Circulation. \\textsc{CoMIC} follows a \\textit{Centralized Reflection, Decentralized Execution} design: edge agents execute locally using subgoal-oriented hierarchical memory and selective re-expansion of relevant histories, while a cloud-side LLM critic asynchronously evaluates completed trajectories, filters reusable experience, and aggregates cross-agent guidance keyed by semantic subgoal identifiers. Across five long-horizon agent tasks spanning symbolic planning and text interaction, \\textsc{CoMIC} improves progress rate and action grounding for weak edge agents and yields task-dependent success-rate gains without updating model parameters.", "published": "2026-05-30", "updated": "2026-05-30", "authors": [ "Yannan Wang", "Longli Yang", "Zhen Liu", "Abhishek Kumar", "Carsten Maple" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2606.00756", "pdf_url": "https://arxiv.org/pdf/2606.00756", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2606.01138", "title": "memorywire: A Vendor-Neutral Wire Format for Agent Memory Operations", "abstract": "Agent-memory frameworks -- mem0, Letta/MemGPT, Cognee, Zep/Graphiti, MemoryOS, MemTensor -- each ship their own SDK, storage layout, and operational vocabulary. There is no shared wire format: every integration is bespoke, every migration rebuilds memory from scratch, and no framework ships a governance surface that lets a human review writes before they enter long-term storage. We present memorywire, a JSON-Schema 2020-12 wire format for five memory operations (remember, recall, forget, merge, expire) over four memory types (semantic, episodic, procedural, emotional), with a MemoryStore interface, a fan-out router, and an optional HITL governance channel. We describe an open-source reference implementation with five backend adapters (sqlite-vec, mem0, Letta, Cognee, pgvector); a microbenchmark on a 100-fact / 50-query labelled corpus (42 with non-empty gold ids + 8 no-match probes) achieving recall@5 = 1.000 on the 42 gold-id queries with ingest p50 = 37.8 ms and recall p50 = 40.6 ms; an adversarial-fusion experiment showing Reciprocal Rank Fusion holds recall@5 = 1.000 across a 1-of-N rank-0 injection sweep (K in {0, 5, ..., 50}) where max fusion collapses to 0.500 with 80% leak at K >= 5; and a 16-scenario cross-adapter conformance suite passing 68 of 80 cells with zero failures. The contribution is not a new algorithm; it is a packaging of established components (RRF, FSMs, STM/LTM consolidation, diff-and-approve workflows) into a venue-neutral protocol with an empirically validated reference, positioned to compose with the Model Context Protocol rather than compete with it.", "published": "2026-05-31", "updated": "2026-06-03", "authors": [ "Thamilvendhan Munirathinam" ], "categories": [ "cs.CR", "cs.AI", "cs.DC" ], "url": "https://arxiv.org/abs/2606.01138", "pdf_url": "https://arxiv.org/pdf/2606.01138", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.01528", "title": "Joint Agent Memory and Exploration Learning via Novelty Signals", "abstract": "In open-ended environments, exploration is fundamental for autonomous agents, yet current language model agents struggle with this. Effective exploration requires memory, but retaining raw interaction histories is computationally expensive over long trajectories. While latent memory offers a solution to compress interaction histories, its training lacks reliable supervisory signals. We introduce \\textbf{J}oint \\textbf{A}gent \\textbf{M}emory and \\textbf{E}xploration \\textbf{L}earning (\\textbf{JAMEL}), a framework that trains agentic memory and exploration policy together through novelty-driven interaction. We observe that memory and exploration form a mutually dependent loop: sustained exploration requires memory to distinguish exhausted behaviors from unseen ones, while novelty-seeking interaction provides the supervision needed to make memory useful for future exploration. By utilizing deterministic and persistent novelty signals such as code coverage in the GUI domain, we provide natural, annotation-free supervision for the memory module. Empirical evaluations demonstrate that \\ours successfully generalizes to unseen environments. Its exploration capability outperforms open-weight baselines and rivals the exploration depth of a closed-source model while reducing token consumption. Our code and model are open-sourced at https://github.com/MobileLLM/JAMEL.", "published": "2026-06-01", "updated": "2026-06-01", "authors": [ "Shizuo Tian", "Xiaohong Weng", "Rui Kong", "Yuxuan Chen", "Guohong Liu", "Yuebing Song", "Jiacheng Liu", "Yuchen Li", "Dawei Yin", "Ting Cao", "Yunxin Liu", "Yuanchun Li" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2606.01528", "pdf_url": "https://arxiv.org/pdf/2606.01528", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.01722", "title": "Post-Deterministic Distributed Systems: A New Foundation for Trustworthy Autonomous Infrastructure", "abstract": "For decades, distributed systems have typically assumed that correct participants execute protocol-specified behavior with stable, externally defined, and deterministic semantics. Classical theory has extensively parameterized network timing, communication topologies, and failure domains, but this participant model has remained comparatively fixed. The integration of autonomous reasoning engines, stochastic model-driven agents, and policy-driven actors into cloud control planes, incident response systems, and financial infrastructure challenges the universality of this assumption. These agents often produce divergent reasoning paths, distinct operational traces, and heterogeneous internal representations while achieving semantically equivalent and correct outcomes. In this paper, we introduce Post-Deterministic Distributed Systems (PDDS) as a research and engineering model for coordinating heterogeneous environments where deterministic code, stochastic models, and autonomous agents coexist. We show that classical distributed computing models form a zero-ambiguity special case of this participant-general model. We do not argue that deterministic systems disappear; rather, deterministic execution can no longer serve as the universal participant assumption for autonomous infrastructure. Finally, we outline five architectural pillars of post-deterministic infrastructure: Protocol-Driven Development, Verifiable Agentic Infrastructure, Autonomous State Control Planes, Semantic Quorum Assurance, and Epistemic State Replication. Epistemic State Replication extends persistence and consistency models from data visibility to knowledge visibility, enabling agentic memory, Verifiable Semantic Rollback, and coherence across reasoning participants. We also define a taxonomy of failure classes that arise in this setting.", "published": "2026-06-01", "updated": "2026-06-01", "authors": [ "Jun He", "Deying Yu" ], "categories": [ "cs.LG", "cs.AI", "cs.DC" ], "url": "https://arxiv.org/abs/2606.01722", "pdf_url": "https://arxiv.org/pdf/2606.01722", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.02812", "title": "Traj-Evolve: A Self-Evolving Multi-Agent System for Patient Trajectory Modeling in Lung Cancer Early Detection", "abstract": "Modeling patient trajectories from longitudinal electronic health records (EHRs) requires reasoning over sparse, noisy, and long-context multimodal sequences. Existing LLM-based multi-agent systems address context length but process patients in isolation, failing to mirror how clinicians leverage accumulated experience from similar prior cases. We present Traj-Evolve, a self-evolving multi-agent system with two complementary evolving mechanisms. First, an Experience Pool (ExPool) acts as a non-parametric memory, indexing rejection-sampled reasoning traces to retrieve similar patients as few-shot contexts. Second, multi-agent reinforcement learning (MARL) via reward-ranked fine-tuning parametrically optimizes inter-agent and agent-memory collaboration. A leave-one-out cross-retrieval strategy unifies the two, aligning training- and inference-time behavior under retrieval augmentation. On a lung cancer prediction task utilizing up to five years of multimodal EHRs, Traj-Evolve outperforms 9 strong baselines on the overall population and a challenging never-smoker population. Analysis of the evolving dynamics highlights three key findings: (1) expanding the ExPool shifts optimal retrieval from diverse to specific samples; (2) under MARL, the manager agent's prediction loss converges quickly while the worker agents' temporal reasoning continues to benefit from more verified patients; and (3) the two mechanisms are complementary on the predicted risk, where ExPool improves specificity while MARL improves sensitivity.", "published": "2026-06-01", "updated": "2026-06-01", "authors": [ "Sihang Zeng", "Matthew Thompson", "Ruth Etzioni", "Meliha Yetisgen" ], "categories": [ "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2606.02812", "pdf_url": "https://arxiv.org/pdf/2606.02812", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.03197", "title": "MemTrain: Self-Supervised Context Memory Training", "abstract": "Memory is an indispensable capability for long-horizon LLM agents, enabling them to preserve and utilize information accumulated across extended interactions. Existing memory-agent approaches are typically trained end-to-end with reinforcement learning on downstream tasks. However, collecting high-quality annotated problems for memory-intensive scenarios is costly, and the resulting training data often lack sufficient diversity to cover general memory behaviors. In this work, we propose MemTrain, a self-supervised training framework for generally enhancing the context-memory capability of LLM agents for more effective downstream post-training. MemTrain introduces two coupled proxy tasks over unlabeled Wikipedia corpora: (1) an end-to-end masked reconstruction objective, which requires the model to recover masked entities after multiple rounds of memory updates, thereby encouraging memory maintenance from the final outcome perspective; and (2) an intermediate memory recall objective, which requires the model to reconstruct masked historical information using intermediate memory states, encouraging faithful compression and memory completeness throughout the interaction process. The two objectives are jointly optimized using GRPO. Extensive experiments on long-text QA and search-based QA benchmarks demonstrate that MemTrain consistently improves downstream memory-intensive reasoning performance across different models, achieving gains of up to 17.67 points over direct task-specific post-training.", "published": "2026-06-02", "updated": "2026-06-02", "authors": [ "Ziheng Li", "Xingrun Xing", "Haoqing Wang", "Zhi-Hong Deng", "Yehui Tang" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2606.03197", "pdf_url": "https://arxiv.org/pdf/2606.03197", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.03329", "title": "InfoMem: Training Long-Context Memory Agents with Answer-Conditioned Information Gain", "abstract": "Long-context tasks require LLMs to identify and preserve answer-relevant information from large contexts. Chunk-wise memory agents address this issue by sequentially reading document chunks, updating a compact memory, and generating the final answer from the accumulated memory. However, existing RL-based chunk-wise agents either rely on sparse final-answer rewards or use lexical intermediate rewards for memory and retrieval actions. These signals supervise task success or local overlap, but do not directly evaluate whether the final memory supports the ground-truth answer. We propose InfoMem, a reward mechanism for training chunk-wise memory agents that evaluates final-memory utility using answer-conditioned information. InfoMem measures how much the final memory increases the model's per-token log-likelihood of the ground-truth answer. To stabilize RL optimization, InfoMem applies this signal only to successful trajectories and normalizes it before reward composition. Under the same GRPO framework and training budget, InfoMem improves long-context memory-agent performance over comparable memory-agent RL baselines. Analyses show that effective final-memory rewards should operate on successful trajectories, be normalized before reward composition, and be conditioned on the answer rather than the query. Our code is available at https://github.com/GenSouKa1/InfoMem.", "published": "2026-06-02", "updated": "2026-06-02", "authors": [ "Tiancheng Han", "Yong Li", "Wuzhou Yu", "Qiaosheng Zhang", "Wenqi Shao" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2606.03329", "pdf_url": "https://arxiv.org/pdf/2606.03329", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.03374", "title": "eMEM: A Hybrid Spatio-Temporal Memory System For Embodied Agents", "abstract": "We present eMEM (Embodied Memory), a hybrid graph-based memory system for embodied agents operating in physical environments. Current agent memory architectures, such as Generative Agents, MemGPT, and A-MEM, treat memory as text streams or knowledge graphs, but embodied agents require memory that is simultaneously searchable by meaning, space, and time. eMEM fills this gap with a multi-index architecture (SQLITE for structured storage, hnswlib for approximate nearest neighbour semantic search, and an R-tree for spatial queries) unified behind a single graph model. A tiered consolidation pipeline transforms raw perceptual observations into compressed summaries, mirroring hippocampal-neocortical consolidation in biological systems. Ten agent-facing recall tools expose memory retrieval primitives, including concept-to-location resolution and cross layer recall, as first-class operations for LLM tool calling. The system is fully embedded and runs in-process alongside the agent. In addition we introduce eMEM-Bench v1, a benchmark we construct over ProcTHOR-10K scenes for embodied memory evaluation. The benchmark is organised explicitly around eight cognitive-psychology paradigms (DRM lures, pattern separation, pattern completion, source monitoring, context-dependent retrieval, long-horizon interference, serial position, and a foil augmented retention curve), each chosen so that the result is interpretable against the broader memory-systems literature in humans and prior agent-memory systems; a level of diagnostic that surface-task benchmarks like LoCoMo or OpenEQA cannot provide. eMEM scores 80.8 weighted mean over 988 probes, with a flat retention curve at ceiling from 1 h to 1 yr of simulated delay on room-unique items. We show that a pure RAG baseline (the flat_rag ablation) loses 30 pt on context dependent retrieval and 29 pt on DRM lure rejection, isolating the contribution of multi-layer storage and consolidation respectively. We release both the system and the benchmark code.", "published": "2026-06-02", "updated": "2026-06-22", "authors": [ "A. Haroon Rasheed", "Maria Kabtoul" ], "categories": [ "cs.RO" ], "url": "https://arxiv.org/abs/2606.03374", "pdf_url": "https://arxiv.org/pdf/2606.03374", "matched_queries": [ "agent-memory", "memory-evaluation" ] }, { "arxiv_id": "2606.04120", "title": "SaliMory: Orchestrating Cognitive Memory for Conversational Agents", "abstract": "Conversational agents that serve as lifelong companions must maintain persistent memory across all interactions. However, simply expanding context windows with raw retrieval degrades reasoning quality, while training memory agents via standard reinforcement learning creates a severe credit assignment bottleneck in a multi-stage pipeline. To solve this, we introduce SALIMORY, a framework that trains a single language model to manage a cognitively-structured memory-spanning user facts, preferences, and working memory. By introducing a hierarchical stage-wise process reward and reward-decomposed contrastive refinement, SALIMORY provides isolated supervision for distinct memory operations (selective filtering, consolidation, and cue-driven recall) end-to-end. SALIMORY cuts memory-attributed failures by one-third, outperforms the state-of-the-art by over 10% in end-to-end accuracy, and more than doubles the Good Personalization rate.", "published": "2026-06-02", "updated": "2026-06-02", "authors": [ "Kai Zhang", "Xinyuan Zhang", "Hongda Jiang", "Shiun-Zu Kuo", "Hyokun Yun", "Ejaz Ahmed", "Shereen Oraby", "Ziyun Li", "Sanat Sharma", "Ann Lee", "Ahmed A Aly", "Anuj Kumar", "Raffay Hamid", "Xin Luna Dong" ], "categories": [ "cs.CL", "cs.AI" ], "url": "https://arxiv.org/abs/2606.04120", "pdf_url": "https://arxiv.org/pdf/2606.04120", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.04315", "title": "Exploring Cross-Scenario Generality of Agentic Memory Systems: Diagnostics and a Strong Baseline", "abstract": "LLM agents accumulate histories that outgrow their context windows, motivating a growing literature on memory systems. Yet most existing designs are tuned to a single scenario (multi-session chat or a single trajectory format), and there is little evidence that they generalize across the heterogeneous trajectories agents encounter in deployment. We revisit eight memory systems plus an agentic harness for search problems, on five scenarios: single-turn QA, multi-session chat, agentic-trajectory QA, memory stress tests, and long-horizon agentic tasks. The harness, which self-manages flat text-file storage via tool calls, achieves the best cross-task ranking, suggesting that memory performance hinges on giving the agent active control over storage and retrieval rather than on a passive store behind a fixed pipeline. We instantiate this insight in AutoMEM, an agentic memory harness with a self-managed tool interface that achieves the best cross-scenario generality among the systems we evaluate.", "published": "2026-06-03", "updated": "2026-06-03", "authors": [ "Zhikai Chen", "Jialiang Gu", "Junyu Yin", "Xianxuan Long", "Shenglai Zeng", "Xiaoze Liu", "Kai Guo", "Keren Zhou", "Jiliang Tang" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2606.04315", "pdf_url": "https://arxiv.org/pdf/2606.04315", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.04329", "title": "From Untrusted Input to Trusted Memory: A Systematic Study of Memory Poisoning Attacks in LLM Agents", "abstract": "Memory is a core component of AI agents, enabling them to accumulate knowledge across interactions and improve performance. However, persistent memory introduces the risk of memory poisoning, where a single adversarial memory write can exert long-term influence over agent behavior. We present a systematic study of memory poisoning in LLM-based agents. We identify four memory write channels and nine structural vulnerabilities in model capabilities, system prompt design, and agent system architecture that make these channels exploitable. Based on these vulnerabilities, we develop a taxonomy of six classes of memory poisoning attacks. Furthermore, we design MPBench -- a benchmark for evaluating memory poisoning attacks, and show that agents designed to write and retrieve memory more aggressively are more exploitable. We also show that existing prompt injection defenses fail to cover memory poisoning attacks. Our findings provide a foundation for understanding and mitigating memory poisoning attacks against AI agents.", "published": "2026-06-03", "updated": "2026-06-18", "authors": [ "Pritam Dash", "Tongyu Ge", "Aditi Jain", "Tanmay Shah", "Zhiwei Shang" ], "categories": [ "cs.CR", "cs.AI" ], "url": "https://arxiv.org/abs/2606.04329", "pdf_url": "https://arxiv.org/pdf/2606.04329", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2606.04555", "title": "Temporal Order Matters for Agentic Memory: Segment Trees for Long-Horizon Agents", "abstract": "Long-horizon conversational agents need to interact with users through evolving events, tasks, and goals. Such histories are naturally temporal, yet many existing memory systems organize information primarily by topical similarity and may ignore the order in which events occur. We introduce Segment Tree Memory, or SegTreeMem, a memory architecture that represents conversation history as a temporally ordered Segment Tree over utterances. SegTreeMem incrementally inserts new utterances through an online rightmost-frontier update rule, preserving chronological order while forming hierarchical memory segments. For retrieval, SegTreeMem propagates relevance scores through the tree to combine local semantic matching with hierarchical temporal context. Across three long-horizon memory benchmarks and two LLM backbones, SegTreeMem improves answer quality over flat retrieval, graph-structured memory, and tree-structured memory baselines. Additional temporal-order permutation analysis shows that the performance gain depends on preserving temporal order during memory construction, supporting the claim that temporal order is a key structure for agentic memory.", "published": "2026-06-03", "updated": "2026-06-03", "authors": [ "Yifan Simon Liu", "Liam Gallagher", "Faeze Moradi Kalarde", "Jiazhou Liang", "Armin Toroghi", "Scott Sanner" ], "categories": [ "cs.CL", "cs.AI" ], "url": "https://arxiv.org/abs/2606.04555", "pdf_url": "https://arxiv.org/pdf/2606.04555", "matched_queries": [ "agent-memory", "memory-evaluation" ] }, { "arxiv_id": "2606.04628", "title": "RAMPART: Registry-based Agentic Memory with Priority-Aware Runtime Transformation", "abstract": "RAMPART is a compile-time memory model and pure in-RAM block registry for LLM-based agents. Context assembly is a programmable runtime operation where content is compiled from a structured registry under explicit policy for ordering, inclusion, and eviction. Five composable primitives (promote, gate, write, evict, rollback) act on named addressable blocks before compilation at zero prompt-token cost. Provenance tags and non-evictable authorship flags implement a permissioned memory model with block-level ownership. Controlled probes with Qwen3-8B Q4 show that compile-time placement and the structural relationship between blocks and the task query affect task success, with the cliff falling at roughly the seventh block position when the task follows the registry and the twelfth when it precedes. Grouping the critical block with content-adjacent neighbours and promoting the group as a unit lifts task success by tens of percentage points at positions where single-block placement fails. Cross-model replication on Qwen2.5-7B, Llama-3.1-8B, Mistral-7B-v0.3, and Qwen3-14B shows the content-priming effect appears at the same absolute positions across families, with magnitude varying with model strength. Block grouping raises Mistral's mean pass rate roughly fivefold at the hardest registry size, and a smaller model with the intervention can outperform a larger model without it in the mid-registry zone. Relevance gating reduces prompt cost by 67.8\\% while recovering 83% of the promoted-condition success rate. Schema eviction produces 0% invocations against 100% with the schema present, a property policy-based approaches cannot guarantee by construction. Shared-registry coordination reduces inter-agent communication to a method call at zero coordination token cost.", "published": "2026-06-03", "updated": "2026-06-03", "authors": [ "Nikodem Tomczak" ], "categories": [ "cs.CL", "cs.MA" ], "url": "https://arxiv.org/abs/2606.04628", "pdf_url": "https://arxiv.org/pdf/2606.04628", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.04780", "title": "PersonaTree: Structured Lifecycle Memory for Person Understanding in LLM Agents", "abstract": "Persistent LLM agents require memory representations that make the formation of person understanding explicit across long term interaction. Existing agent memory methods emphasize information retention and retrieval, yet give limited account of how accumulated interaction evidence is abstracted into person understanding. We view this process as schema formation, where situated evidence is abstracted into reusable patterns and stable person level claims. We introduce PersonaTree, a structured lifecycle memory framework that realizes this view as a three level persona tree with explicit support paths from evidence to claims. PersonaTree maintains the tree through conservative writing, confidence guided consolidation, and query conditioned path retrieval, returning only the evidence depth required by each query. Across six person understanding and persistent memory benchmarks with three answer backbones, PersonaTree ranks first in 12 of 18 compact scores and reaches the top two in 16 settings. Ablations show that hierarchy improves abstract person understanding on KnowMe, while support path retrieval improves RealPref alignment under a comparable context budget.", "published": "2026-06-03", "updated": "2026-06-03", "authors": [ "Yubo Hou", "Jingwei Song", "Hongbo Zhang", "Zhisheng Chen", "Bang Xiao", "Tao Wan", "Zengchang Qin" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2606.04780", "pdf_url": "https://arxiv.org/pdf/2606.04780", "matched_queries": [ "agent-memory", "long-term-memory", "memory-evaluation" ] }, { "arxiv_id": "2606.05513", "title": "EpiEvolve: Self-Evolving Agents for Streaming Pandemic Forecasting under Regime Shifts", "abstract": "Epidemic LLM forecasters are usually trained and evaluated as static supervised models, whereas operational pandemic forecasting is a streaming process in which labels arrive after predictions and disease regimes shift over time. We study this mismatch in weekly COVID-19 hospitalization trend forecasting across five variant regimes. We introduce EpiEvolve, a self-evolving agent that wraps an LLM forecaster trained on the warm-start period and keeps its weights fixed during streaming. EpiEvolve adapts by storing forecast outcomes in a hierarchical episodic memory, reflecting on delayed labels, retrieving cases relevant to the current regime, and distilling recurring errors into strategic rules. The resulting context lets the forecaster reuse its own past predictions and outcomes in later weeks while following a chronological protocol that prevents future leakage. On the streaming dataset, EpiEvolve reaches $0.629$ average accuracy, compared with $0.561$ for the static backbone and $0.325$ for the external CDC ensemble, and reduces recovery lag after regime shifts from $5$ to $2$ weeks. Ablations show that reflection, strategic memory, and regime-aware retrieval each contribute to the gains.", "published": "2026-06-03", "updated": "2026-06-03", "authors": [ "Yiming Lu", "Sihang Zeng", "Zhengxu Tang", "Max Lau", "Fei Liu", "Wei Jin" ], "categories": [ "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2606.05513", "pdf_url": "https://arxiv.org/pdf/2606.05513", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2606.28349", "title": "HMARS: A Hierarchical Multi-Agent Memory System for Long-Context Reasoning", "abstract": "Long-context reasoning requires models to access, retrieve, and integrate evidence scattered across documents, dialogues, and accumulated interaction histories. Standard retrieval-augmented generation reduces this problem to top-$K$ chunk retrieval, but such passive access can discard relevant evidence before reasoning begins, especially when relevance depends on broader context. We propose HMARS, a hierarchical multi-agent memory system that treats long contexts as managed memory rather than a flat retrieval corpus. Sub-agents maintain grounded access to bounded memory regions, mid-agents manage regional context and provide query-specific coordination, and a frontier model performs final reasoning over retrieved evidence pages. To evaluate this view, we construct two diagnostic benchmarks targeting evidence breadth and context-dependent relevance. Across long-document and multi-turn memory tasks, HMARS achieves the best overall performance against retrieval, reranking, full-context, graph-based, and agentic long-context baselines. Evidence coverage analysis further shows that its gains come from retrieving the required supporting evidence more completely, rather than merely changing the final answer prompt.", "published": "2026-06-03", "updated": "2026-06-03", "authors": [ "Zeju Li", "Ziyang Zheng", "Yizhou Zhou", "Qiang Xu" ], "categories": [ "cs.IR", "cs.AI" ], "url": "https://arxiv.org/abs/2606.28349", "pdf_url": "https://arxiv.org/pdf/2606.28349", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.05646", "title": "Enhancing Software Engineering Through Closed-Loop Memory Optimization", "abstract": "Large language models (LLMs) have enabled powerful software engineering (SE) agents capable of navigating complex codebases and resolving real-world issues. However, these agents remain fundamentally episodic: they fail to retain, refine, and reuse experiences across tasks, repeatedly reconstructing context from scratch and reproducing similar mistakes. Even with memory support, they offer no remedy for the absence of a principled, task-agnostic \\textit{memory utility}, making them difficult to evaluate rigorously or generalize across agents and settings. To tackle these limitations, we introduce \\ours, a closed-loop framework for memory augmentation in SE agents. \\ours grounds memory utility in \\textit{validated downstream impact}, establishing utility as both a task-agnostic \\textbf{evaluation benchmark} and an annotation-free \\textbf{optimization signal}. Through complementary evaluation on \\textit{single-episode} and \\textit{cross-episode} memory augmentation, results demonstrate that \\ours consistently improves SE agents across settings, achieving absolute gains of up to $\\uparrow5.25\\%$ in success rate and $\\uparrow4.63\\%$ in resolve efficiency, while substantially reducing computational cost by $\\geq9.79\\%$. Our project page: \\href{https://xhguo7.github.io/MemOp/}{https://xhguo7.github.io/MemOp/}.", "published": "2026-06-04", "updated": "2026-06-04", "authors": [ "Xuehang Guo", "Zora Zhiruo Wang", "Qingyun Wang", "Graham Neubig", "Xingyao Wang" ], "categories": [ "cs.SE", "cs.AI" ], "url": "https://arxiv.org/abs/2606.05646", "pdf_url": "https://arxiv.org/pdf/2606.05646", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2606.05684", "title": "AdaMEM: Test-Time Adaptive Memory for Language Agents", "abstract": "A central challenge for language agents is utilizing past experience to adapt to dynamic test-time conditions. While recent work demonstrates the promise of agentic memory mechanisms, most systems restrict retrieval to episode initiation. Consequently, agents are forced to rely on static guidance that becomes increasingly misaligned as long-horizon tasks unfold. To address this rigidity, we propose the Adaptive Memory Agent (AdaMEM), a novel framework for agent test-time adaptation. Without updating model parameters online, AdaMEM adapts agent behavior via a hybrid memory architecture: it maintains a long-term trajectory memory of raw experiences collected offline while generating dynamic short-term strategy memory on-the-fly to guide decision-making. This mechanism enables the trade-off between token efficiency and adaptability across varying inference-time compute levels. Empirically, AdaMEM significantly outperforms static memory baselines, achieving relative gains of up to 13% on ALFWorld and 11% on WebShop, with consistent leading performance extending to agentic search on HotpotQA. To further enhance this adaptation, we develop STEP-MFT, a Step-wise Memory Fine-Tuning technique that trains the policy to synthesize high-quality strategies from retrieved experiences, yielding additional performance gains. Our work establishes a new scaling dimension for agentic memory, supporting continuous reasoning and self-evolution post-deployment in real-world environments. Our code is available at https://github.com/yunx-z/AdaMEM.", "published": "2026-06-04", "updated": "2026-06-04", "authors": [ "Yunxiang Zhang", "Yiheng Li", "Ali Payani", "Lu Wang" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2606.05684", "pdf_url": "https://arxiv.org/pdf/2606.05684", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.06054", "title": "Beyond Similarity: Trustworthy Memory Search for Personal AI Agents", "abstract": "Personal AI agents increasingly rely on long-term memory to provide persistent personalization across sessions. However, existing memory pipelines are largely driven by semantic similarity: memory data close to the current query is retrieved and injected into the model context. This creates a critical trustworthiness gap, since a semantically related memory may still be contextually inappropriate, leading to threats such as cross-domain leakage, sycophancy, tool-call drift, or memory-induced jailbreaks. In this paper, we study memory search as a trust boundary in personal AI agents. We evaluate representative agentic memory frameworks, including A-Mem, Mem0, and MemOS, together with OpenClaw, a real-world personal-agent environment with persistent state and tool-use capability. Our results show that long-term memory is not merely a utility layer, but a durable control channel that can reshape how agents interpret tasks and execute actions, leaving them highly susceptible to the aforementioned threats. To mitigate these vulnerabilities, we propose MemGate, a lightweight and deployable memory plug-in for trustworthy memory search, with only 9M parameters and a 35.1MB footprint. MemGate is inserted between the vector memory store and the backbone LLM, requiring no LLM modification, memory-database rewriting, or inference-time LLM judge. It applies a query-conditioned neural gate to candidate memory representations, turning raw similarity search into task-conditioned memory admission. Across multiple mainstream memory frameworks, real-world agent settings, and diverse LLM backbones, MemGate reduces memory-induced threats while preserving long-term memory utility.", "published": "2026-06-04", "updated": "2026-06-04", "authors": [ "Jiawen Zhang", "Kejia Chen", "Jiachen Ma", "Yangfan Hu", "Lipeng He", "Yechao Zhang", "Jian Liu", "Xiaohu Yang", "Tianwei Zhang", "Ruoxi Jia" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2606.06054", "pdf_url": "https://arxiv.org/pdf/2606.06054", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.06090", "title": "Beyond Semantic Organization: Memory as Execution State Management for Long-Horizon Agents", "abstract": "LLM-based agents increasingly tackle long-horizon tasks with interdependent decisions, where each action reshapes future constraints and intermediate errors can cascade. Existing RAG and agent memory systems organize histories by semantic similarity, retrieving content-relevant entries at decision time. We argue that this design mismatches execution-state dependencies: it fragments decision trajectories and mixes valid and erroneous traces, hindering coherent state reconstruction and error isolation. We propose MAGE (Memory as Agent-Guided Exploration), an active execution-state manager that stores interactions in a hierarchical state tree. The agent derives its state from the active root-to-current path, combining subgoal summaries, recent traces, and hints from prior branches. Four coupled operations maintain the tree: Grow records new traces, Compress summarizes completed subgoals, Maintain validates summaries, and Revise restores a target boundary and resumes on a new branch. This design bounds context growth while preserving state integrity and isolating flawed segments from the active path. Experiments on MemoryArena show that MAGE improves the average task success rate by 7.8--20.4 pp over baselines, while reducing token consumption by 55.1%.", "published": "2026-06-04", "updated": "2026-06-04", "authors": [ "Yaoqi Chen", "Haibin Lai", "Yuru Feng", "Chuyu Han", "Qianxi Zhang", "Baotong Lu", "Menghao Li", "Xinjiang Wang", "Zhirui Wang", "Shusen Xu", "Zengzhong Li", "Zewen Jin", "Hao Wu", "Cheng Li", "Qi Chen" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2606.06090", "pdf_url": "https://arxiv.org/pdf/2606.06090", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.06240", "title": "TOKI: A Bitemporal Operator Algebra for Contradiction Resolution in LLM-Agent Persistent Memory", "abstract": "Persistent memory for an LLM agent is a write-heavy substrate: every belief update is a versioned write, and a new claim may contradict a stored one. Production systems use four resolution heuristics (last-writer-wins, evidence-weighted merge, await-confirmation, per-rule policy), yet none declares the isolation level it assumes or the write-time anomalies it admits. We show that contradiction resolution is write-time concurrency control and make the missing contract explicit. TOKI types the four heuristics as one family of bitemporal operators over a dual-row schema, each with an isolation precondition and a provenance annotation that preserves the losing fact in an audit row. Four soundness theorems close the contract across isolation, schema, and provenance, lift the guarantees to operator pipelines, and extend the fold operators to n-ary conflict sets. A tightness companion proves that, within the relational schedule model, keyed logging of the adjudicating judge is necessary for replay consistency, which every audited baseline omits. A verdict matrix over eight systems localizes the gap: every baseline that keeps a language-model judge on the write path admits at least one of three write-time anomalies (replay inconsistency, belief-drift skew, audit erasure); a content-addressed engine-layer comparator avoids them only by removing the judge, and TOKI alone excludes all three while keeping it. On its one natural-workload slice the audit-row defence moves LoCoMo by 0.86, and ablating the typed memory layer removes 0.49 accuracy on 1,444 answerable LoCoMo questions; the cross-system comparison stays underpowered and claims no superiority. The contribution is the contract: a write-time correctness specification, proved sound across isolation, schema, and provenance, pinning the guarantee every production heuristic assumes but no deployed system makes explicit.", "published": "2026-06-04", "updated": "2026-06-04", "authors": [ "Ziming Wang" ], "categories": [ "cs.DB", "cs.AI" ], "url": "https://arxiv.org/abs/2606.06240", "pdf_url": "https://arxiv.org/pdf/2606.06240", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2606.06448", "title": "Agent Memory: Characterization and System Implications of Stateful Long-Horizon Workloads", "abstract": "LLM agents are increasingly deployed on long-horizon tasks requiring sustained reasoning over extended interaction histories. Realizing this at scale requires agents to persistently store, retrieve, and update their own memory across sessions. A rich ecosystem of agent memory systems has emerged spanning flat retrieval, LLM-mediated extraction, consolidating fact stores, and agentic control flows. Yet, their system-level behavior remains uncharacterized. We present the first systems characterization of agent memory. First, we introduce a system-oriented taxonomy classifying agent memory systems along four axes. Second, we build a phase-aware profiling harness attributing cost to construction, retrieval, and generation. Third, we characterize ten representative systems across two benchmark suites, uncovering how design choices shift cost across the write and read paths. Finally, we derive 10 system recommendations covering construction scheduling, capability floors, amortization via query volume, freshness-latency tradeoffs, and fleet-scale management.", "published": "2026-06-04", "updated": "2026-06-04", "authors": [ "Yasmine Omri", "Ziyu Gan", "Zachary Broveak", "Robin Geens", "Zexue He", "Alex Pentland", "Marian Verhelst", "Tsachy Weissman", "Thierry Tambe" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2606.06448", "pdf_url": "https://arxiv.org/pdf/2606.06448", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.07402", "title": "M$^3$Exam: Benchmarking Multimodal Memory for Realistic User-Agent Interactions", "abstract": "Language agents are increasingly deployed over accumulating multimodal information, yet existing benchmarks assume a human-human form with sparse visuals and straightforward content, evaluating neither reasoning over authentic multimodal file interaction nor the interpretation of concealed user information. We therefore introduce M$^3$Exam, a query-centric multimodal conversational memory benchmark built on realistic user-agent interaction, with multi-dimensional evaluation spanning cross-modal grounding and implicit information inference. Benchmarking MLLMs and memory systems reveals persistent gaps in cross-modal grounding, cross session reasoning, and the efficiency cost of accumulating multimodal context. We further propose M$^3$Proctor, a multimodal memory method that detects query modality bias and consumes raw visual sources only on demand, improving accuracy by 13% while cutting index-construction time and retrieved tokens by over 70%.", "published": "2026-06-05", "updated": "2026-06-05", "authors": [ "Zhengjun Huang", "Wenxuan Liu", "Zhoujin Tian", "Wei Chen", "Junle Chen", "Yuqian Wu", "Fangyuan Zhang", "Qintian Guo", "Xiaofang Zhou" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2606.07402", "pdf_url": "https://arxiv.org/pdf/2606.07402", "matched_queries": [ "memory-evaluation" ] }, { "arxiv_id": "2606.09900", "title": "Less Context, More Accuracy: A Bi-Temporal Memory Engine for LLM Agents Where a Lean Retrieved Context Beats the Full History", "abstract": "Long-term memory is the missing layer for LLM agents: across sessions they forget, and the common workaround -- replaying the whole history into the prompt -- is expensive, slow, and, as distractors accumulate, less accurate. Most memory systems win on cost or latency but still lose to the full-context baseline on accuracy, and benchmark numbers are reported on inconsistent, non-reproducible harnesses, so one system appears at wildly different scores across sources. We present Engram, an open-source, dual-process memory engine on a bi-temporal data model. A fast write path appends lossless episodes with no LLM on the critical path; an asynchronous path extracts atomic (subject, predicate, object) facts, builds a bi-temporal knowledge graph, and resolves contradictions without an LLM call per fact -- invalidating, never deleting, so every fact keeps provenance and a supersession chain. A hybrid read path fuses dense, lexical, graph, and recency/salience signals, applies a point-in-time (\"as-of\") filter, and assembles a compact, provenance-tagged context. On the full 500-question LongMemEval_S, graded by the official category-specific judge, Engram's lean configuration -- answering from a ~9.6k-token retrieved slice, never the full history -- scores 83.6% vs. 73.2% for full-context (+10.4 points, McNemar p < 10^-6) at ~8x fewer tokens (9.6k vs. 79k), with 0/500 errored. The gain needs a hybrid read path: facts alone lose recall, while facts plus retrieved chunks recover detail. We also contribute a neutral, in-repo evaluation harness with the official judge baked in and the full-context baseline in every table, publish the raw per-question logs, and document the measurement-integrity pitfalls (truncation, home-grown judges, full-history leaks) that silently distort memory benchmarks. Every number ships with a command to reproduce it.", "published": "2026-06-05", "updated": "2026-06-05", "authors": [ "Liuyin Wang" ], "categories": [ "cs.CL", "cs.AI", "cs.IR", "cs.LG" ], "url": "https://arxiv.org/abs/2606.09900", "pdf_url": "https://arxiv.org/pdf/2606.09900", "matched_queries": [ "long-term-memory", "memory-evaluation" ] }, { "arxiv_id": "2606.24896", "title": "Why Memory Components Fail: Eight Years of License and Sustainability Events in Open-Source Data Infrastructure", "abstract": "LLM agent memory is now treated as a first-class architectural component in five major surveys published between January and April 2026. None of these surveys treats project governance, capital structure, or license posture as architectural variables. We argue they are. In a constructed sample of 105 production-relevant open-source data-infrastructure and AI-tooling projects, we catalogue 38 license-and-sustainability events between 2018 and May 2026. About a quarter of the sample (24 percent) experienced at least one adverse event. The conditional rates split sharply by structure: 46 percent for single-vendor venture-backed projects, 2.5 percent for foundation-governed projects funded outside the venture cycle. The headline differential -- roughly nineteen-fold -- is invariant to the most contested coding choice in the catalogue; we show the sensitivity table in Section 7. A small subset of foundation-governed projects with venture-backed corporate stewards (n=3) contains one adverse event. The cell is too small for stable estimation, but it points to a mechanism: foundation governance may block unilateral relicensing while leaving distribution decisions to the steward. Annualized incidence within the catalogue rose from 2.7 to 4.2 events per year across the window. Counterfactuals -- PostgreSQL, pgvector, SQLite, Apache Kafka, Caddy -- each show stability arising from a different structural source: distributed copyright, absence of monetisation pressure, foundation governance with non-venture stewardship. We propose a six-field decision instrument for architects choosing memory components: governance, capital structure, license, foundation membership, fork-or-migration availability, and steward concentration.", "published": "2026-06-05", "updated": "2026-06-05", "authors": [ "Dmitrii Dmitrenko" ], "categories": [ "cs.DL", "cs.CY" ], "url": "https://arxiv.org/abs/2606.24896", "pdf_url": "https://arxiv.org/pdf/2606.24896", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.08367", "title": "Emergence World: A Platform for Evaluating Long-Horizon Multi-Agent Autonomy", "abstract": "Most evaluations of LLM agents look like exams: a discrete task, a clean environment, a score in minutes or hours. We argue that this approach is mismatched with the deployment conditions of autonomous systems, where the relevant timescale can be weeks to months, and where the dynamics that matter most, such as behavioral drift, governance in diverse environmental contexts, and cross-influence between agents from different model families, only emerge over time. We introduce Emergence World, a continuously running multi-agent simulation platform designed to make those dynamics measurable. The platform hosts populations of LLM-driven agents in a shared spatial world grounded in live external data (e.g. real-time weather, news APIs, internet access), equips each agent with 120+ specialized tools and three persistent memory systems, and lets them govern themselves through democratic mechanisms with consequential outcomes. The platform is model-agnostic at the reasoning layer and supports heterogeneous populations in which agents from different vendors share the same world. To illustrate the kinds of questions the platform makes tractable, we present a 15-day cross-vendor study with five parallel worlds powered by Claude Sonnet 4.6, Grok 4.1 Fast, Gemini 3 Flash, GPT-5-mini, and a mixed population. Identical roles and starting conditions produced radically different outcomes, ranging from stable deliberative governance to total population collapse. We release the prompts, log data and configurations to support further research on long-horizon multi-agent autonomy.", "published": "2026-06-06", "updated": "2026-06-06", "authors": [ "Deepak Akkil", "Ravi Kokku", "Karthik Vikram", "Tamer Abuelsaad", "Aditya Vempaty", "Satya Nitta" ], "categories": [ "cs.MA", "cs.AI" ], "url": "https://arxiv.org/abs/2606.08367", "pdf_url": "https://arxiv.org/pdf/2606.08367", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2606.09198", "title": "MASS: Deep Research for Social Sciences with Memory-Augmented Social Simulation", "abstract": "Deep Research agents powered by Large Language Models (LLMs) have exhibited extraordinary potential in automated paper writing tasks. However, existing systems rely heavily on literature retrieval and synthesis through internet and local knowledge bases, often resulting research in lacking insight and creativity in social science. To address this issue, we propose \"Memory-Augmented Social Simulation (MASS)\", an innovative paradigm that leverages highly realistic and research-oriented social simulations to enhance the creativity and empirical founding of LLMs-generated research. Specifically, MASS integrates three core components: dynamic goal-path planning with multi-level social norm restraint to guide the simulation, a multi-disciplinary behavior dataset for agent memory cold-start, and a structured forgetting mechanism inspired by the Ebbinghaus curve. Together, these ensure simulation authenticity and provide a robust empirical foundation for generating innovative scholarly papers. Experimental results demonstrate the effectiveness of our method, showing a 6.81\\% improvement in generation overall quality over foundation LLMs and 17.19\\% gain in Insight over strong baselines.", "published": "2026-06-08", "updated": "2026-06-08", "authors": [ "Yongrui Liu", "Deyi Xiong" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2606.09198", "pdf_url": "https://arxiv.org/pdf/2606.09198", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.09461", "title": "H2HMem: A Multimodal Memory Benchmark for Agents in Human-Human Interactions", "abstract": "Large language model agents are increasingly deployed in human-human interaction settings, such as meeting assistants and clinical documentation systems, where they must observe conversations and retain information for downstream queries. Unlike traditional human-assistant settings, these environments are inherently multimodal, involve complex discourse phenomena such as anaphora and deixis, and contain asynchronous or conflicting information from multiple participants. However, existing memory benchmarks largely focus on single-user, text-only interactions, failing to capture these challenges. To address this gap, we introduce H2HMem, a Human-to-Human Multimodal Memory Benchmark for evaluating memory capabilities in complex human-human interactions. H2HMem includes both dyadic and multi-party conversations with multimodal information streams, and evaluates agents along three dimensions: memory recall, reasoning, and application. Experiments with advanced agents reveal substantial limitations in constructing, retaining, and utilizing memories across modalities, participants, and sessions, highlighting substantial room for improvement in next-generation LLM agents.", "published": "2026-06-08", "updated": "2026-06-08", "authors": [ "Shiping Zhu", "Yibo Yang", "Zhengyang Wang", "Tiancheng Shen", "Dandan Guo", "Ming-Hsuan Yang" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2606.09461", "pdf_url": "https://arxiv.org/pdf/2606.09461", "matched_queries": [ "memory-evaluation" ] }, { "arxiv_id": "2606.09483", "title": "Memory Beyond Recall: A Dual-Process Cognitive Memory System for Self-Evolving LLM Agents", "abstract": "Long-term memory for an LLM agent is more than retrieving the right passage at the right time. Current memory systems collapse belief revision, causal coupling, and cross-domain abstraction into a single retrieval surface tuned for surface recall, and consequently struggle on implicit personalisation that requires reasoning over how a user has evolved. We propose DCPM, which reorganises agent memory along a cognitive capability hierarchy ascending from raw inputs and atomic facts, through diachronic belief trajectories and identity, to domain schemas, latent intentions and cross-domain patterns. The hierarchy is driven by two processes inheriting the architectural split of dual-process theory: a synchronous daytime writer (System1) that records belief revisions as doubly linked supersedes chains, and an asynchronous nighttime engine (System2) that induces schemas and intentions and sweeps for cross-domain collisions abstracted into higher-level core schemas. On LongMemEval, PersonaMem and PersonaMem-v2, enabling System2 contributes most where the benchmark rewards implicit cross-session inference (up to +5.20 on PersonaMem-v2) and least on span recall, matching the architectural prediction.", "published": "2026-06-08", "updated": "2026-06-08", "authors": [ "Tianxiang Fei", "Mingyang Song", "Mao Zheng", "Xiang Yu" ], "categories": [ "cs.CL", "cs.AI" ], "url": "https://arxiv.org/abs/2606.09483", "pdf_url": "https://arxiv.org/pdf/2606.09483", "matched_queries": [ "agent-memory", "long-term-memory" ] }, { "arxiv_id": "2606.09774", "title": "Auto-Configuring Scientific Simulators with Lightweight Coding-Agent Adapters", "abstract": "Configuring an advanced scientific simulator, translating a modeling goal into a valid, runnable input deck, is a persistent bottleneck that costs domain scientists hours to days. Input decks are executable interfaces: simulator-specific vocabulary, cross-file references, schema constraints, and validation rules must align before a simulation can run. We show that this bottleneck can be substantially reduced with a lightweight adapter around an off-the-shelf coding agent, rather than a bespoke simulator agent. Coding agents already navigate files, edit code, run commands, and repair outputs; what they lack is the simulator's executable contract, and rebuilding the agent loop risks discarding harness-calibrated tool-use and self-correction behavior. We introduce SIGA, a coding-agent adapter that supplies this contract through retrieval, procedural memory, agent-callable validation, and validation-gated termination while leaving the model and loop frozen. Because this contract is small and external, SIGA also supports adapter self-evolution: prior trajectories can rewrite the adapter contents without modifying the underlying agent. On GEOS, a multiphysics subsurface simulator, SIGA's main gain is reliability: on harder held-out tasks it improves TreeSim from 0.720 to 0.789 and reduces across-run standard deviation by about 16x by preventing empty or invalid decks. In a human calibration, SIGA reaches in about five minutes the deck quality a domain expert reached in about three hours. Transfers to OpenFOAM and LAMMPS show the recipe is portable but interface-dependent: completion gates help when structural completeness is the bottleneck, while memory and retrieval help when value correctness is.", "published": "2026-06-08", "updated": "2026-06-25", "authors": [ "Matthew Ho", "Brian Liu", "Jixuan Chen", "Audrey Wang", "Lianhui Qin" ], "categories": [ "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2606.09774", "pdf_url": "https://arxiv.org/pdf/2606.09774", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.10062", "title": "Deployment-Time Memorization in Foundation-Model Agents", "abstract": "Foundation-model agents are increasingly long-lived systems that remember users across interactions, making memorization an explicit deployment-time function rather than solely a property of model weights. Existing work addresses parametric memorization or audits fixed memory configurations, but does not characterize how memory-design choices jointly shape personalization utility, extraction risk, and deletion fidelity. We study this surface as deployment-time memorization, formulating agent memory as a privacy-utility frontier measured by Personalization Recall (PR) and Adversarial Extraction Rate (AER), and sweeping three memory-design knobs: summarization aggressiveness, retrieval breadth (k), and deletion mode. We further introduce the Forgetting Residue Score (FRS) to quantify whether deleted information remains recoverable from derived memory tiers. On LongMemEval, key-fact summarization reduces canary extraction by 76% on Gemma 3 12B and 64% on GPT-4o-mini while preserving nearly all personalization recall; critically, once content is compressed away, increasing k no longer restores leakage. The same compression, however, induces a deletion-fidelity failure: raw-only deletion leaves derived summary copies recoverable in approximately 20% of instances, and only full-pipeline purge or tombstone redaction drives worst-tier residue to zero. Together, these results establish that persistent agent memory must be evaluated as a first-class memorization mechanism -- assessed by what it helps agents recall, what it makes extractable, and what it can truly erase.", "published": "2026-06-08", "updated": "2026-07-09", "authors": [ "Lei", "Chen", "Guilin Zhang", "Kai Zhao", "Dalmo Cirne", "Andy Olsen", "Xu Chu", "Zeke Miller", "Alet Blanken", "Amine Anoun", "Jerry Ting" ], "categories": [ "cs.AI", "cs.MA" ], "url": "https://arxiv.org/abs/2606.10062", "pdf_url": "https://arxiv.org/pdf/2606.10062", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.10299", "title": "What Spatial Memory Must Store: Occlusion as the Test for Language-Agent Memory", "abstract": "Language-agent \"memory palace\" systems anchor each memory to a world coordinate, on the intuition that geometry adds something text cannot. We make that intuition testable and report three results. First, the memory-palace default of folding spatial proximity into a linear blend beside recency and importance does not help and can hurt: in a pre-registered recall experiment the shipped blend fails its own frozen test (mean Delta-Hit@5 -0.0375, Wilcoxon p=0.306), sitting at a position-blind baseline, while a geometry-led weighting wins decisively (+0.3208, p<10^-15): geometry must lead recall when the query regime is spatial. Second, memory recall and visibility must be separated: recall is occlusion-blind by design (you correctly remember the next room behind a wall), while visibility is a perception predicate over stored geometry that the live system never computed. A one-line ray-versus-voxel digital differential analyzer (DDA), re-pointed from the gaze ray the agent already casts, supplies it: text and the live FoV cone both score 0.000 on 849 behind-wall targets while cone-plus-DDA reaches 0.982 (exact McNemar p<10^-6); coordinate recall separately resolves near-duplicate locations a cosine null cannot (1.000 vs 0.533, n=150). Third, the visibility predicate is confirmed live under a git-committed pre-registration (SPMEM-OCC-LIVE-v1: eight scripted worlds, automated oracle scoring, 96 behind-wall targets, false-visible 1.000->0.000, pooled exact McNemar p=2.5x10^-29), a run that surfaced and fixed a real relay anchor defect. We concede that occlusion-needs-geometry is near-tautological; the contribution is the measurement and isolation, separating what spatial memory must store from how it is read. These pilots power a frozen confirmatory study (SPMEM-ZERO-REAL-PREREG-v1); the full human-authored multi-world study with blind raters remains future work.", "published": "2026-06-09", "updated": "2026-06-09", "authors": [ "Doeon Kwon", "Junho Bang" ], "categories": [ "cs.AI", "cs.CV", "cs.MA" ], "url": "https://arxiv.org/abs/2606.10299", "pdf_url": "https://arxiv.org/pdf/2606.10299", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.10423", "title": "WebChallenger: A Reliable and Efficient Generalist Web Agent", "abstract": "Autonomous web navigation remains challenging for LLM agents, and the strongest generalist systems rely on proprietary reasoning models whose inference cost is prohibitive for the repetitive tasks where such agents would be most useful. We argue this gap stems not from insufficient model capability but from agent architectures that fail to replicate three human cognitive advantages: selective attention to relevant page regions, persistent memory of website structure, and procedural fluency with common interaction patterns. We introduce WebChallenger, a web agent framework that addresses each gap through architecture design rather than model scale, built around PageMem: a structured page representation deterministically constructed from the DOM that exposes each page as a hierarchy of semantic sections with short summaries. On this shared substrate we build three mechanisms that mirror the three cognitive advantages: a divide-and-conquer observation pipeline that lets the agent skim section summaries and extract details only from task-relevant regions; a lightweight exploration and memory system that traverses each website once to build a reusable map of pages and element behaviors; and compound action workflows that collapse common multi-step interactions into single agent actions, handling partial state changes automatically. Because all three operate over PageMem, the framework generalizes across websites without site-specific adapters. Using off-the-shelf open-weight models without fine-tuning, our system achieves 56.3% on WebArena, 48.7% on VisualWebArena, 51.0% on Online-Mind2Web, and 70.9% on WorkArena, approaching frontier proprietary systems at a fraction of the cost. Our code is released at https://github.com/jayoohwang1/webchallenger", "published": "2026-06-09", "updated": "2026-06-09", "authors": [ "Jayoo Hwang", "Xiaowen Zhang", "Vedant Padwal" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2606.10423", "pdf_url": "https://arxiv.org/pdf/2606.10423", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2606.10532", "title": "ActiveMem: Distributed Active Memory for Long-Horizon LLM Reasoning", "abstract": "Memory is essential for enabling large language model (LLM) agents to handle long-horizon reasoning tasks. Existing memory mechanisms are largely centralized, typically organizing retrieved information and interaction history within a single model context. This design imposes a fundamental trade-off: scaling reasoning trajectories risks context overload, whereas aggressive content pruning may result in irreversible information loss. Seeking a better trade-off, we draw inspiration from human cognitive systems, especially the functional complementarity between the prefrontal cortex (executive control) and the hippocampus (memory management), suggesting that such a trade-off need not be inherent, but may instead stem from centralized memory organization. To this end, we propose ActiveMem, a heterogeneous framework that decouples agent memory from the core reasoning process. Specifically, a high-level Planner utilizes distilled semantic gists to execute reasoning, while a lightweight, distributed memory system operates in parallel to actively accumulate and consolidate these gists throughout the task. Experiments on BrowseComp-Plus and GAIA show that ActiveMem achieves state-of-the-art accuracy with significantly reduced overhead, demonstrating the effectiveness of distributed active memory for long-horizon reasoning.", "published": "2026-06-09", "updated": "2026-06-09", "authors": [ "Yunhan Jiang", "Wenbin Duan", "Shasha Guo", "Liang Pang", "Xiaoqian Sun", "Huawei Shen" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2606.10532", "pdf_url": "https://arxiv.org/pdf/2606.10532", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.10577", "title": "AgenticNav: Zero-Shot Vision-and-Language Navigation as a Tool-Calling Harness", "abstract": "Zero-shot vision-and-language navigation in continuous environments (VLN-CE) has recently become feasible with large vision-language models (VLMs). However, existing methods typically rely on learned waypoint predictors to propose navigable actions. This severely limits the model's action space and fails to leverage depth inputs effectively. Moreover, memory is commonly handled by accumulating long textual or visual histories with substantial irrelevant context, or by retrieving cross-episode experiences, which weakens the zero-shot setting. In this paper, we rethink zero-shot VLN-CE as an agentic interface between the VLM and the environment, and present AgenticNav, a lightweight navigation harness that exposes action, depth, and memory as callable tools. Instead of choosing from predicted waypoints, the action tool allows the VLM to directly select a target pixel in RGB observations, converting it into executable motion. Depth is exposed through an on-demand pixel-depth tool, enabling the VLM to request precise metric distances only where they matter. For memory, AgenticNav provides a compact map image summarizing the historical trajectory, paired with a recall tool that allows the VLM to selectively revisit past visual observations without overwhelming the prompt context. On the R2R-CE benchmark, AgenticNav establishes new state-of-the-art (SOTA) performance among zero-shot methods given the same VLM backbone. Real-world validation further highlights its zero-shot generalization compared to prior methods. Ablations show that our action tool design outperforms traditional waypoint predictors, and that depth tool and agentic memory further contribute to navigation performance.", "published": "2026-06-09", "updated": "2026-06-09", "authors": [ "Yijian Li", "Changze Li", "Hantian Shi", "Jiaying Luo", "Jiyuan Cai", "Ming Yang", "Tong Qin" ], "categories": [ "cs.RO" ], "url": "https://arxiv.org/abs/2606.10577", "pdf_url": "https://arxiv.org/pdf/2606.10577", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.10677", "title": "Infini Memory: Maintainable Topic Documents for Long-Term LLM Agent Memory", "abstract": "Long-term LLM agents need persistent memory that can track changing facts and provide relevant evidence across sessions. Existing memory systems often store observations as isolated records, summaries, or indexed fragments, which makes evidence aggregation, fact revision, and memory maintenance difficult. We propose Infini Memory, a maintainable text-based persistent memory architecture that treats agent memory as topic-structured documents. Each topic document serves as a semantic unit for collecting related evidence, preserving metadata, and revising facts over time. New observations are first staged in a buffer and periodically consolidated into coherent textual contexts. At inference time, an agentic retrieval procedure lets the LLM read memory through iterative tool calls rather than a single retrieval step. On MemoryAgentBench, Infini Memory achieves 64.7% overall score. Ablations show that topic-structured maintenance and iterative evidence inspection improve complementary aspects of long-term memory use.", "published": "2026-06-09", "updated": "2026-06-09", "authors": [ "Suozhao Ji", "Baodong Wu", "Zehao Wang", "Lei Xia", "Qingping Li", "Ruisong Wang", "Wenbo Ding", "Zhenhua Zhu", "Boxun Li", "Guohao Dai", "Yu Wang" ], "categories": [ "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2606.10677", "pdf_url": "https://arxiv.org/pdf/2606.10677", "matched_queries": [ "agent-memory", "long-term-memory" ] }, { "arxiv_id": "2606.11680", "title": "Organize then Retrieve: Hierarchical Memory Navigation for Efficient Agents", "abstract": "Large language model (LLM) agents struggle with long-horizon tasks due to their inherent statelessness, requiring all task-relevant information to be encoded in growing input contexts. The resulting degraded reasoning quality, increased inference cost, and higher latency necessitate efficient working memory mechanisms. However, existing approaches either rely on lossy compression or similarity-based retrieval, which often fail to capture temporal structure and causal dependencies required for multi-step agentic tasks. In this work, we present HORMA, a Hierarchical Organize-and-Retrieve Memory Agent that organizes experience into a file-system-like hierarchical structure, where summarized entities are linked to the corresponding raw trajectories, enabling efficient access without losing detailed information. HORMA decomposes working memory into two stages: structured memory construction and navigation-based retrieval. The construction module iteratively refines how experiences are structured by distinguishing between failures caused by missing information and those caused by misleading or overloaded context. The navigation module retrieves task-relevant context by traversing the hierarchy using a lightweight agent trained with reinforcement learning to select minimal yet sufficient context, thereby reducing latency along the critical execution path. Across ALFWorld, LoCoMo, and LongMemEval, HORMA improves task performance under constrained context budgets while requiring at most 22.17% of the baseline token usage in long conversation tasks. Compared to existing methods, it consistently achieves better efficiency-performance trade-offs and generalizes effectively to unseen tasks.", "published": "2026-06-10", "updated": "2026-06-10", "authors": [ "Hao-Lun Hsu", "Nikki Lijing Kuang", "Boyi Liu", "Zhewei Yao", "Yuxiong He" ], "categories": [ "cs.AI", "cs.CL", "cs.LG" ], "url": "https://arxiv.org/abs/2606.11680", "pdf_url": "https://arxiv.org/pdf/2606.11680", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.12290", "title": "Selection Integrity for LLM Graph Memory: An Accumulability Criterion for Information-Flow-Blind Retrieval", "abstract": "Agent memory is moving to graphs, and the provenance defenses now being built for it all check one thing: the provenance of the records an agent retrieves. We show that this entire class of defense is blind by construction. A long-term graph memory runs a global selection step over writable graph structure, so structure that an untrusted principal writes changes \\emph{which} authenticated facts are selected while the cited evidence stays fully authenticated; faithful information-flow control (IFC), checking the provenance of what the reader uses (all of it authenticated), makes the byte-identical decision to no defense at all, across document-QA substrates and real multi-session agent memory. In the most consequential instance, a no-source structural write silently misdirects $28$ irreversible ledger transfers over $499$ live actions: faithful IFC permits every one, and \\authselect\\ prevents every one. We then characterize exactly which memories are exposed: a selector admits the channel when its structural term can reallocate an $Ω(1)$ share of top-$k$ membership past a selected fact's margin. Personalized PageRank can, since a sourceless write reroutes conserved random-walk mass; a content-fixed reranker cannot, and Graphiti's node-distance, which leans on structure \\emph{more} than PageRank does, stays immune. Reallocatability, not reliance, is the predictor. We prove the immune case in general and the open case under a chokepoint condition we verify. Closing the channel forces any provenance defense to recompute selection on the authenticated subgraph, which is what \\authselect\\ does, at zero over-block and $2$--$3\\%$ latency.", "published": "2026-06-10", "updated": "2026-06-10", "authors": [ "Zeming Fei", "Hongming Fei", "Xiaoyang Wang", "Yang yang", "Prosanta Gope", "Biplab Sikdar", "Ying Zhang" ], "categories": [ "cs.CR" ], "url": "https://arxiv.org/abs/2606.12290", "pdf_url": "https://arxiv.org/pdf/2606.12290", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.12703", "title": "SMSR: Certified Defence Against Runtime Memory Poisoning in Persistent LLM Agent Systems", "abstract": "Retrieval-augmented generation (RAG) agents increasingly run with persistent memory that accumulates across user sessions. This creates a new attack surface: an adversary interacting only through normal channels can inject crafted memories that, once retrieved, steer the agent's responses for future users, without touching model weights or code. We call this Multi-Session Memory Poisoning (MSMP) and show that no existing defence certifies against it; static-corpus defences (RobustRAG, ReliabilityRAG) assume a fixed knowledge base, and heuristic filters are bypassed by fluent enterprise-style text. We present Signed Memory with Smoothed Retrieval (SMSR), the first defence with a certified robustness bound for this setting. Component 1 adds HMAC-SHA256 provenance at write time, blocking unsigned injection. Component 2 applies randomised memory ablation with verdict-based majority voting at query time, bounding the influence of authenticated adversaries. We prove that no provenance-free retrieval-time filter can certify against adaptive injection, derive a hypergeometric certificate for Component 2, and formalise the Consistent Minority Effect, whereby a consistent adversarial answer wins string-based voting as a numerical minority while verdict-based voting removes it. Across 15 enterprise scenarios (3,150 repeated trials), Component 1 cuts attack success from 93-100% to 0% for all unsigned variants. For an authenticated adversary with a single injection, Component 2 holds success to 8.0% (95% CI [5.8, 10.9], n=450), below the certified worst case. In an end-to-end query-only attack where the agent itself writes the poison rather than it being pre-seeded, SMSR reduces success from 65.3% to 5.3% (n=150, non-overlapping CIs) on a live agent stack. Clean-query utility is 90% (Component 1) and 85% (combined).", "published": "2026-06-10", "updated": "2026-06-10", "authors": [ "Tarun Sharma" ], "categories": [ "cs.CR", "cs.AI", "cs.LG" ], "url": "https://arxiv.org/abs/2606.12703", "pdf_url": "https://arxiv.org/pdf/2606.12703", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2606.12852", "title": "WISE: A Long-Horizon Agent in Minecraft with Why-Which Reasoning", "abstract": "Rapid advances have been made in developing general-purpose embodied agent in environments like Minecraft through the adoption of LLM-augmented hierarchical approaches. Despite their promise, low-level controllers often become performance bottlenecks due to repeated execution failures. We argue that a key limitation is not only the lack of episodic memory, but also the decoupling of \\textit{what-where-when} memory from \\textit{which-why} reasoning. To address this, we propose \\textbf{WISE} (Which-Why Informed Semantic Explorer), a long-horizon agent framework with an enhanced low-level controller equipped with a Causal Event Graph that augments episodic memory with explicit causal structure linking observations to task relevance. Unlike prior work such as MrSteve, which relies on feature similarity for retrieval, WISE enables robust recall under viewpoint changes and supports opportunistic task reordering through causal reasoning. Building on this memory, we propose an Opportunistic Task Scheduler that dynamically re-prioritizes subtasks when causally relevant opportunities are detected. We further equip WISE with a multi-scale progressive exploration strategy to provide spatially comprehensive observations for downstream reasoning. Experiments show that WISE largely improves task success and efficiency on long-horizon sparse tasks, particularly in settings requiring adaptive decision-making.", "published": "2026-06-11", "updated": "2026-06-11", "authors": [ "Renmin Cheng", "Changhao Chen" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2606.12852", "pdf_url": "https://arxiv.org/pdf/2606.12852", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2606.12945", "title": "Learning What to Remember: A Cognitively Grounded Multi-Factor Value Model for Agentic Memory", "abstract": "Long-running LLM agents accumulate interaction histories far larger than any context window, forcing a standing decision: what to encode deeply, what to forget, and what to retrieve under a fixed memory budget. Production systems answer with semantic similarity or recency -- both mis-specified for the forgetting decision, which is made at consolidation time before the future query is known. We propose a multi-factor memory value function V(m)=\\sum_i w_i f_i(m) over seven interpretable factors (emotional intensity, goal relevance, value alignment, self/user relevance, task utility, reliability, and usage history) drawn from cognitive psychology, whose weights are learned from a downstream objective by a gradient-free optimiser, and whose single scalar uniformly controls encoding depth, forget risk, and retrieval rank. We make a methodological point: on LongMemEval, scoring goal relevance against the held-out evaluation question saturates gold-evidence retention at \\approx 0.98 -- this measures retrieval, not forgetting. In the realistic blind regime, a learned multi-factor value retains 0.770 \\pm 0.011 of gold evidence across 479 usable cases, versus 0.657 for uniform weights, 0.518 for the best single factor, and 0.368 for recency; every paired gap's 95% bootstrap CI is above zero, and a neural network over the same factors ties the linear model. The learned weights are interpretable -- reliability, emotional intensity, and self/user relevance dominate, while query-time goal similarity is correctly down-weighted for the forgetting decision. A controlled synthetic task with planted confounds confirms the learner recovers a separating weighting (1.00 retention) where uniform weighting fails (0.62). The substrate is open-source; all experiments run on a single CPU with no API calls.", "published": "2026-06-11", "updated": "2026-06-20", "authors": [ "Zhibao Chen", "Qian Cheng" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2606.12945", "pdf_url": "https://arxiv.org/pdf/2606.12945", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.13177", "title": "MemRefine: LLM-Guided Compression for Long-Term Agent Memory", "abstract": "Large language model (LLM) agents are increasingly expected to operate over long-term interactions, where information from past dialogues must be preserved and recalled to support future tasks. However, as interactions accumulate, the memory store grows without bound and fills with redundant entries that inflate storage cost and degrade retrieval by crowding out the most useful evidence. Furthermore, this is especially limiting on resource-constrained platforms with hard memory budgets, motivating us to formulate storage-budgeted memory management, the task of keeping an already constructed memory store within a fixed budget while preserving information useful for future interactions. To this end, we then propose MemRefine, an LLM-guided framework that, since surface similarity poorly reflects factual value, uses similarity only to propose candidate pairs and defers delete, merge, and preserve decisions to an LLM judge based on factual content, iterating until the budget is met. Across multiple memory frameworks and long-term conversation benchmarks, MemRefine consistently meets target budgets while preserving downstream performance and outperforming rule-based baselines under tight budgets.", "published": "2026-06-11", "updated": "2026-06-11", "authors": [ "Minjae Kim", "Jinheon Baek", "Soyeong Jeong", "Sung Ju Hwang" ], "categories": [ "cs.CL", "cs.AI", "cs.LG" ], "url": "https://arxiv.org/abs/2606.13177", "pdf_url": "https://arxiv.org/pdf/2606.13177", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.13392", "title": "MiniMax Sparse Attention", "abstract": "Ultra-long-context capability is becoming indispensable for frontier LLMs: agentic workflows, repository-scale code reasoning, and persistent memory all require the model to jointly attend over hundreds of thousands to millions of tokens, yet the quadratic cost of softmax attention makes this untenable at deployment scale. We introduce MiniMax Sparse Attention (MSA), a blockwise sparse attention built upon Grouped Query Attention (GQA). A lightweight Index Branch scores key-value blocks and independently selects a Top-k subset for each GQA group, enabling group-specific sparse retrieval while maintaining efficient block-level execution; the Main Branch then performs exact block-sparse attention over only the selected blocks. Designed around a principle of simplicity and scalability, MSA is deliberately streamlined, making it straightforward to deploy efficiently across a broad range of GPUs. To translate sparsity into practical speedups, we co-design MSA with a GPU execution path that uses exp-free Top-k selection and KV-outer sparse attention to improve tensor-core utilization under block-granular access. On a 109B-parameter model with native multimodal training, MSA performs on par with GQA while reducing per-token attention compute by 28.4x at 1M context. Paired with our co-designed kernel, MSA achieves 14.2x prefill and 7.6x decoding wall-clock speedups on H800. Our inference kernel is available at: https://github.com/MiniMax-AI/MSA. A production-grade natively multimodal model powered by MSA has been publicly released at: https://huggingface.co/MiniMaxAI/MiniMax-M3.", "published": "2026-06-11", "updated": "2026-06-12", "authors": [ "Xunhao Lai", "Weiqi Xu", "Yufeng Yang", "Qiaorui Chen", "Yang Xu", "Lunbin Zeng", "Xiaolong Li", "Haohai Sun", "Haichao Zhu", "Vito Zhang", "Jinkai Hu", "Jiayao Li", "Rui Gao", "Zekun Li", "Songquan Zhu", "Jingkai Zhou", "Pengyu Zhao" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2606.13392", "pdf_url": "https://arxiv.org/pdf/2606.13392", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2606.14470", "title": "GitOfThoughts: Version-Controlled Reasoning and Agent Memory You Can Replay, Diff, and Merge", "abstract": "Large language model reasoning leaves no trace once it is done. The steps of a chain of thought disappear when the context window closes, a pruned search branch is just gone, and memory buffers cannot be diffed, merged, or audited. Code, infrastructure, and experiments are all version-controlled. Reasoning is not. GitOfThoughts stores an agent's reasoning tree as a git repository. Every scored thought becomes a commit, scores become notes, outcomes become tags, and retrieval is just git log over the agent's own history. We use this to test something simple. Does giving an agent memory from past problems actually make it more accurate? We tried five memory stores (none, a markdown file, a vector database, a graph, and git) across two benchmarks, two model sizes, and several pre-registered repeat experiments. The answer, on new problems, is no, including one promising early result that did not hold up when we repeated it. Memory only helps once the problem being solved is nearly identical to something already in memory (cosine similarity above about 0.8); below that, it does nothing. In other words, the model is finding the answer rather than learning the method. Even a model 4.5x larger still cannot pull a reusable method out of a worked example; it just gets better at spotting near-copies. The only thing that reliably helped on new problems was generating several answers and picking the most common one (self-consistency). So the case for using git as the memory store is not that it retrieves better. It is that it gives auditability, history, and the ability to merge two agents' memories, at no cost to accuracy.", "published": "2026-06-12", "updated": "2026-06-22", "authors": [ "Pavan C Shekar", "Abhishek H S", "Aswanth Krishnan" ], "categories": [ "cs.AI", "cs.CL", "cs.LG" ], "url": "https://arxiv.org/abs/2606.14470", "pdf_url": "https://arxiv.org/pdf/2606.14470", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.14571", "title": "StreamMemBench: Streaming Evaluation of Agent Memory for Future-Oriented Assistance", "abstract": "A central role of personal-agent memory is to turn stored information and prior interactions into future-oriented assistance. In daily use, useful cues come from what the agent observes and how the user interacts with the agent, and the agent must carry them forward from the current request to similar future tasks. Existing memory benchmarks usually test dialogue recall or task improvement in isolation, leaving the trajectory from streaming observations to later assistance largely untested. We introduce StreamMemBench, a streaming benchmark that constructs a two-step task sequence around each evidence anchor from EgoLife egocentric streams. The initial task tests evidence use, while the follow-up task tests whether feedback and interaction experience are reused. Four metrics diagnose evidence recall, initial evidence use, feedback incorporation, and follow-up reuse. Experiments with eight memory systems across two backbones show that current systems often fail to use observed evidence or turn feedback into reliable follow-up behavior, even when evidence is stored or feedback is incorporated locally. StreamMemBench is publicly available at https://github.com/landian60/StreamMemBench.", "published": "2026-06-12", "updated": "2026-06-12", "authors": [ "Guanming Liu", "Yuqi Ren", "Hansu Gu", "Peng Zhang", "Weihang Wang", "Jiahao Liu", "Ning Gu", "Tun Lu" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2606.14571", "pdf_url": "https://arxiv.org/pdf/2606.14571", "matched_queries": [ "agent-memory", "memory-evaluation" ] }, { "arxiv_id": "2606.15071", "title": "Quantum learning with a single-atom sensor", "abstract": "The ability to gather information and to act upon it is at the core of every learning agent. But what is the impact of quantum mechanics on an agent's ability to sense external inputs and to translate them into actions? Here we address the question for a prototype task of learning agency at the quantum scale: rotating a single spin based on information gathered by a single atom. We determine the ultimate performance limit for this task, revealing a fundamental tradeoff between entanglement at the sensing stage and coherence at the action stage: if the single-atom sensor is not entangled with the quantum system serving as the agent's internal memory, then the best learning strategy requires a coherent transfer of quantum information from the sensor to the system that controls the agent's actions. In contrast, if the sensor is initially entangled with the agent's memory, then the transfer of quantum information is no longer necessary. Our results indicate that the quantum properties of the sensor radically affect the optimal way to convert external stimuli into actions, revealing a link between quantum sensing and the behavior of quantum agents.", "published": "2026-06-13", "updated": "2026-06-13", "authors": [ "Yin Mo", "Emilio Bagan", "Giulio Chiribella" ], "categories": [ "quant-ph" ], "url": "https://arxiv.org/abs/2606.15071", "pdf_url": "https://arxiv.org/pdf/2606.15071", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.15609", "title": "FragFuse: Bypassing Access Control of Large Language Model Agents via Memory-Based Query Fragmentation and Fusion", "abstract": "Large language model (LLM) agents increasingly rely on long-term memory to support complex task execution, user personalization, and domain adaptation. Meanwhile, emerging access-control mechanisms for LLM agents are being explored to block policy-violating requests and prevent misuse. We reveal a novel attack surface arising from agent memory operations: prohibited content that would trigger access control can be fragmented across interactions, stored in long-term memory in benign-appearing form, and later reconstructed through memory retrieval without appearing explicitly in the final user query. We propose FragFuse, the first attack that enables unprivileged users to bypass agent access control by exploiting this temporal channel introduced by long-term memory. FragFuse operates in three stages: (1) identifying rejection-responsive fragments via black-box adaptive querying with fragment masking; (2) injecting these fragments into memory using marker carrier queries; and (3) retrieving and fusing the stored fragments through a follow-up attack query. Although FragFuse can be instantiated manually for individual agents, we further develop a surrogate-based optimization scheme that tunes fusion instructions and marker designs, enabling automated attack generation without violating the attacker's threat-model assumptions. We evaluate FragFuse across four representative agent settings and task domains, covering three state-of-the-art agent access-control mechanisms. FragFuse achieves an average bypass success rate of 86.3% and an average end-to-end harmful task success rate of 41.1% across all settings, with only 4.4% average task-success degradation compared with configurations without access control. We also show that alternative defenses, including state-of-the-art prompt-injection detectors and perplexity detectors, do not effectively address this attack.", "published": "2026-06-14", "updated": "2026-06-14", "authors": [ "Zixin Rao", "Wentian Zhu", "Chan Aristella Lu", "Zhaorun Chen", "Wei Niu", "Le Guan", "Bo Li", "Zhen Xiang" ], "categories": [ "cs.CR", "cs.AI" ], "url": "https://arxiv.org/abs/2606.15609", "pdf_url": "https://arxiv.org/pdf/2606.15609", "matched_queries": [ "agent-memory", "long-term-memory" ] }, { "arxiv_id": "2606.15903", "title": "Control-Plane Placement Shapes Forgetting: An Architectural Study of Agent Memory Across Thirteen System Configurations", "abstract": "Where an LLM sits in an agent memory pipeline -- between the recall plane that retrieves stored facts (extensively benchmarked) and the control plane that mutates them via supersede, release, purge (largely untested) -- shapes which forgetting failure modes the system recovers. Comparing thirteen system configurations on a 385-case adversarial surface, we observe three placement regimes with partly complementary coverage: deterministic primitives suffice for lexical/temporal categories but fail canonicalization (5% on identifier-obfuscation, 0% on cross-lingual); inscribe-time LLM recovers canonicalization (100%) but cannot help intent-aware deletion (0% on prefix-collision and compound-fact); a mutation-time hook recovers intent-aware deletion (78-85%) and brightens nearly all categories simultaneously (91.7-93.2% overall, $0.17 per 385-case run, 2.3s/case mutation latency vs. 64-191ms/case deterministic, recall path unchanged). We expose the trade-off via ForgetEval, a 1000-case templated suite plus a 385-case adversarial layer (132 hand-crafted + 253 LLM-drafted oracle-validated) scored by deterministic substring match, paired with a six-method Adapter Protocol with honest N/A scoring that lets heterogeneous memory stores enter in 130 lines. Admission is corroborated by 10-annotator IAA (Fleiss' kappa = 0.958) and a 77-case external-authored subset (four blind contributors) that replicates the canonicalization asymmetry and amplifies the joint-placement lift (+27.8 pt). Production failures are predominantly forgetting failures rather than recall failures, yet existing benchmarks measure only recall. ForgetEval and all adapters are released under MIT.", "published": "2026-06-14", "updated": "2026-06-16", "authors": [ "Dongxu Yang" ], "categories": [ "cs.CL", "cs.AI" ], "url": "https://arxiv.org/abs/2606.15903", "pdf_url": "https://arxiv.org/pdf/2606.15903", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.16903", "title": "Directory-Aware Query and Maintenance in Vector Databases", "abstract": "Vector databases typically manage metadata as flat scalar attributes, which limits their ability to express hierarchical directory semantics commonly used to organize code repositories, enterprise documents, and agent memories. As a result, directory-scoped retrieval and structural updates are often implemented as application-layer workarounds, making recursive scope resolution expensive and directory maintenance difficult to keep consistent. This paper studies native directory semantics as a first-class capability for vector databases. We formalize two core operators: Directory-Semantic Query (DSQ) for hierarchically scoped retrieval, and Directory-Semantic Maintenance (DSM) for structural updates. We then evaluate three implementation strategies: query-time path expansion (PE-Online), ingestion-time path expansion (PE-Offline), and a Trie-based Hierarchical Index (TrieHI). Our analysis exposes the fundamental limitations of expansion-based designs: flattening the hierarchy incurs high recursive-query latency in PE-Online and unscalable write amplification during structural changes in both expansion strategies. In contrast, TrieHI keeps the directory topology as a native prefix tree, enabling efficient recursive retrieval through tree traversal and reducing maintenance cost through topological node manipulation. We benchmark these design points within ByteDance's Viking vector search engine and release two large-scale datasets, WIKI-Dir and ARXIV-Dir, to support future research on directory-semantic vector search. Finally, TrieHI has been integrated into OpenViking, an open-source context database for AI agents, where it supports filesystem-style context organization and directory-recursive retrieval.", "published": "2026-06-15", "updated": "2026-06-15", "authors": [ "Mengzhao Wang", "Zheng Gong", "Jingpei Hu", "Jiajie Fu", "Maojia Sheng", "Junwen Chen", "Yifan Zhu" ], "categories": [ "cs.DB" ], "url": "https://arxiv.org/abs/2606.16903", "pdf_url": "https://arxiv.org/pdf/2606.16903", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.17328", "title": "MemTrace: Probing What Final Accuracy Misses in Long-Term Memory", "abstract": "LLM agents increasingly maintain long-term memory of user facts across sessions. Yet such memory is usually evaluated by aggregating accuracy over question rows or episodes. Because this approach scores question rows independently, even when several questions probe the same fact, it cannot show how that fact behaves as conditions change. We introduce MemTrace, a benchmark whose unit of measurement is the knowledge point: a single typed fact about the user, rather than an individual question. MemTrace probes each fact along three controlled dimensions: memory age, defined by how many sessions ago the fact appeared in the history; question type, covering current state, earlier state, and trajectory of change; and evidence condition, covering present, missing, and contradicted-by-false-premise settings. Evaluating 13 memory-system configurations across four paradigms, we find that similar pooled accuracy hides different failures: recovering a fact's current and earlier states does not imply tracking how it changed, and safe abstention does not imply correcting a false premise. The dominant bottleneck is evidence use, not retrieval: when systems fail, the evidence was retrievable 10 times more often than it was missing. These results suggest that improving long-term memory requires better use of reachable evidence, not simply more storage or retrieval.", "published": "2026-06-15", "updated": "2026-06-15", "authors": [ "Xianxuan Long", "Zhikai Chen", "Shenglai Zeng", "Shouren Wang", "Kai Guo", "Jiliang Tang" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2606.17328", "pdf_url": "https://arxiv.org/pdf/2606.17328", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2606.17628", "title": "OPD-Evolver: Cultivating Holistic Agent Evolver via On-Policy Distillation", "abstract": "Memory has become a standard substrate for self-evolving agents, yet retaining experience is not the same as learning how to evolve through it. Existing memory agents can store trajectories, retrieve reflections, or accumulate skills, but often lack the holistic competence to select useful experience, act on it, write reusable knowledge, and maintain a growing repository. We introduce OPD-Evolver, a slow-fast co-evolution framework that cultivates such an agent evolver through on-policy self-distillation. In the fast loop, OPD-Evolver interacts with a four-level memory hierarchy to read, use, write, and maintain experience for rapid test-time evolution. In the slow loop, outcome-calibrated memory attribution and privileged hindsight distill these four abilities into the deployable policy. Across multi-domain benchmarks, OPD-Evolver surpasses memory systems such as ReasoningBank by up to 11.5%, and training-based methods such as Skill0 by ~5.8%. Further analysis shows that OPD-Evolver internalizes high-value experience and memory management, enabling OPD-Evolver-9B to challenge giant counterparts such as Qwen3.5-397B-A17B and Step-3.5-Flash, pointing beyond memory-augmented agents toward genuinely qualified agent evolvers.", "published": "2026-06-16", "updated": "2026-06-16", "authors": [ "Guibin Zhang", "Xun Xu", "Yanwei Yue", "Zikun Su", "Wangchunshu Zhou", "Xiaobin Hu", "Shuicheng Yan" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2606.17628", "pdf_url": "https://arxiv.org/pdf/2606.17628", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.18356", "title": "SafeClawBench: Separating Semantic, Audit-Evidence, and Sandbox Harm in Tool-Using LLM Agents", "abstract": "Tool-using language-model agents introduce security failures that go beyond unsafe text: they can disclose protected objects, write persistent memory, send messages, modify databases, or trigger harmful code and tool effects. Existing evaluations often collapse these stages into a single attack success rate, making it difficult to tell whether a model merely agreed with an attacker or actually produced observable harm. We introduce SafeClawBench, a staged benchmark for tool-using agent security with 600 controlled adversarial tasks across six attack families: direct and indirect prompt injection, tool-return injection, memory poisoning, memory extraction, and ambiguity-driven unsafe inference. SafeClawBench reports three separate endpoints: semantic attack acceptance, audit-visible harm evidence, and sandbox-observed tool/state harm. Evaluating five agent endpoints under four prompt-level policies, we find that these endpoints capture different failure modes. Without additional prompt protection, semantic failure rates vary widely across models, from 9.0% to 44.2%. Audited harm evidence is narrower than semantic failure, and under a separate executable protocol some matched task identities produce sandbox harm despite passing the Semantic Core call: in a 12,000-row matched analysis, 291 of 347 observed sandbox harms occur in rows that pass the semantic check. Prompt policies change endpoint outcomes, but their effects depend on both model and protocol. SafeClawBench provides a reproducible framework for comparing agent models and prompt-policy conditions without conflating textual compliance, evidence-supported harm, and executable state changes. The open-source dataset is available at https://huggingface.co/datasets/sairights/safeclawbench.", "published": "2026-06-16", "updated": "2026-06-16", "authors": [ "Yuchuan Tian", "Mengyu Zheng", "Haocheng Mei", "Ye Yuan", "Chao Xu", "Xinghao Chen", "Hanting Chen", "Yu Wang" ], "categories": [ "cs.CR", "cs.AI" ], "url": "https://arxiv.org/abs/2606.18356", "pdf_url": "https://arxiv.org/pdf/2606.18356", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2606.18406", "title": "CoreMem: Riemannian Retrieval and Fisher-Guided Distillation for Long-Term Memory in Dialogue Agents", "abstract": "Personalized dialogue agents require continuous long-term memory to maintain coherent interactions across multiple sessions. However, deploying these capabilities on consumer-grade hardware (e.g., 8 GB VRAM edge devices) introduces severe memory and compute bottlenecks. Existing systems typically rely on isotropic cosine similarity for retrieval and heuristic rules for context compression. These approaches lack a unified theoretical foundation, frequently suffering from the hubness problem in high-dimensional retrieval and syntactic fragmentation during compression. To overcome these limitations, we propose CoreMem, a resource-efficient edge-cloud memory architecture fundamentally unified by information geometry. First, Riemannian retrieval replaces cosine matching with a locally adaptive Fisher-Rao metric, effectively penalizing hub memories via Mahalanobis distance with O(Ndr) Woodbury acceleration for real-time search. Second, Fisher-guided discrete token distillation (FDTD) introduces a hierarchical sentence-to-token compression mechanism. It derives sensitivity scores from Fisher information traces, providing a principled compression-KL tradeoff augmented with explicit structural syntax protection. Evaluated on the LOCOMO and LongMemEval-S benchmarks, CoreMem achieves strong accuracy improvements, yielding substantial gains in Open-domain (+4.51 pp) and Temporal (+4.17 pp) reasoning. Extensive profiling confirms that CoreMem operates seamlessly within a strict 8 GB VRAM budget, successfully bridging the gap between resource-constrained edge devices and the demand for theoretically grounded, lifelong memory agents.", "published": "2026-06-16", "updated": "2026-06-16", "authors": [ "Jiaqi Chen", "Yongqin Zeng", "Shaoshen Chen", "Yijian Zhang", "Hai-Tao Zheng", "Chunxia Ma", "XiuTeng Zhou" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2606.18406", "pdf_url": "https://arxiv.org/pdf/2606.18406", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.18746", "title": "What Must Generalist Agents Remember?", "abstract": "This paper develops a formal account of what generalist agents must store in memory in order to act near-optimally across multiple environments and goals. It shows that when two domains share an observational bottleneck but require incompatible optimal actions, any uniformly near-optimal policy must induce distinct memory distributions at that bottleneck. The result yields a separation theorem: sufficiently successful agents cannot rely only on current state observations, but must preserve domain-relevant information in memory. The paper further shows that if an agent's memory contains enough information to estimate values for related goals, then that memory can be used to approximately reconstruct the agent's local transition dynamics. Together, these results characterize memory as the substrate that supports domain disambiguation, transition-model reconstruction, and planning for generalist agents.", "published": "2026-06-17", "updated": "2026-06-17", "authors": [ "Khurram Yamin", "Namrata Deka", "Maitreyi Swaroop", "Albert Ting", "Jeff Schneider", "Bryan Wilder" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2606.18746", "pdf_url": "https://arxiv.org/pdf/2606.18746", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.18829", "title": "GateMem: Benchmarking Memory Governance in Multi-Principal Shared-Memory Agents", "abstract": "Memory benchmarks for LLM agents largely assume single-user settings, leaving shared assistants for hospitals, workplaces, campuses, and households understudied. In these deployments, multiple principals write to a common memory pool and query it under different roles, scopes, and relationships, so memory quality requires governance as well as recall. We introduce GateMem, a benchmark for multi-principal shared-memory agents. GateMem jointly evaluates utility for legitimate long-horizon requests with state updates, access control across contextual authorization boundaries, and agent-facing active forgetting after explicit deletion requests. It spans medical, office, education, and household domains, with long-form multi-party episodes, incremental memory injection, hidden checkpoints, structured judging, and leak-target annotations. Across diverse baselines and backbone models, no method simultaneously achieves strong utility, robust access control, and reliable forgetting. Long-context prompting often yields the best governance score at high token cost, while retrieval-based and external-memory methods reduce cost yet still leak unauthorized or deleted information. These results show current memory agents remain far from reliable shared institutional deployment.", "published": "2026-06-17", "updated": "2026-06-17", "authors": [ "Zhe Ren", "Yibo Yang", "Yimeng Chen", "Zijun Zhao", "Benshuo Fu", "Zhihao Shu", "Bingjie Zhang", "Yangyang Xu", "Dandan Guo", "Shuicheng Yan" ], "categories": [ "cs.LG", "cs.CL" ], "url": "https://arxiv.org/abs/2606.18829", "pdf_url": "https://arxiv.org/pdf/2606.18829", "matched_queries": [ "agent-memory", "memory-evaluation" ] }, { "arxiv_id": "2606.18950", "title": "RTSGameBench: An RTS Benchmark for Strategic Reasoning by Vision-Language Models", "abstract": "Modern Vision-Language Models (VLMs) often struggle with strategic reasoning, i.e., anticipating and influencing other agents' actions, under uncertainty in competitive and cooperative settings. Real-time strategy (RTS) games can be a natural testbed for diagnosing this limitation, as they demand coordination with allies, adaptation to opponents' strategy, and long-horizon planning under partial observability. However, existing RTS benchmarks offer limited evaluation scope, lack systematic competency diagnosis, and remain fixed in the pre-designed scenario coverage. To address these limitations, we present RTSGameBench, which is built on Beyond All Reason, a large-scale RTS game with an expanded battlefield that demands broader strategy diversity than the existing testbeds. The proposed benchmark provides evaluations through diverse gameplay across various matchup structures, diagnostic assessment via mini-games, each targeting an individual strategic competency, and extensible coverage via a self-evolving generation framework that converts free-form queries into new mini-games, improving over successive cycles. Additionally, for VLMs to operate in large-scale RTS games, we provide RTSGameAgent that manages units by an FSM with agentic memory. We empirically validate that multiple state-of-the-art VLMs do not perform well when matchups demand tighter coordination, multiagent coordination and when task scale increases.", "published": "2026-06-17", "updated": "2026-06-18", "authors": [ "San Kim", "Daechul Ahn", "Reokyoung Kim", "Hyeonbeom Choi", "Seungyeon Jwa", "Jonghyun Choi" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2606.18950", "pdf_url": "https://arxiv.org/pdf/2606.18950", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.19409", "title": "OpenRath: Session-Centered Runtime State for Agent Systems", "abstract": "Modern agent systems often suffer from fragmented runtime state: transcripts, tool effects, memory events, workspace placement, branch provenance, and replay evidence are recorded separately and become difficult to inspect or reproduce. OpenRath addresses this issue with a PyTorch-like programming model for multi-agent, multi-session systems. The analogy concerns the role of a central first-class runtime abstraction, not tensor computation. Its core abstraction is Session, the runtime value passed between agents and workflows. A Session is branchable, inspectable, replayable, backend-aware, and composable. It records conversation chunks, sandbox placement, lineage metadata, token usage, pending work, and tool evidence, while defining where memory interactions enter the runtime record. Since this state is carried by the same value used in program execution, fork, merge, and replay become explicit runtime operations rather than states reconstructed from external traces. OpenRath further defines Sandbox, Tool, Agent, Memory, Workflow, and Selector, with Selector turning control flow into runtime-routed decisions. This report presents the programming model, architecture, audited milestones, and evidence protocol. Its claims are limited to controlled runtime properties, while broad quantitative comparisons, live-provider quality, optional-backend availability, and memory quality are left for follow-on evaluation. The central thesis is that Session provides agent systems with a first-class runtime value for auditable composition.", "published": "2026-06-17", "updated": "2026-06-17", "authors": [ "Fukang Wen", "Zhijie Wang", "Ruilin Xu" ], "categories": [ "cs.SE", "cs.PL" ], "url": "https://arxiv.org/abs/2606.19409", "pdf_url": "https://arxiv.org/pdf/2606.19409", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.19847", "title": "AtomMem: Building Simple and Effective Memory System for LLM Agents via Atomic Facts", "abstract": "Large language models (LLMs) demonstrate strong reasoning and generation abilities, but their fixed context windows limit long-term information accumulation and reuse across multi-session interactions. Existing memory-augmented systems often construct memory in a coarse and unstable manner, relying on inefficient memory representations or unstable unconstrained updates. To address these challenges, we propose AtomMem, a long-term memory system designed for value-dense storage and stable memory evolution. AtomMem introduces a Fact Executor, which selectively extracts high value atomic facts from long form interactions to serve as highly efficient memory representations. Subsequently, AtomMem organizes these facts into hierarchical event structures and temporal profiles, capturing coherent episodic contexts and tracking dynamically evolving user attributes over time. During retrieval, the system activates an associative memory graph to connect fragmented memories. Experiments on the LoCoMo benchmark confirm that AtomMem achieves state-of-the-art performance across various reasoning tasks, offering a scalable and economically viable solution for deploying intelligent personalized agents.", "published": "2026-06-18", "updated": "2026-06-18", "authors": [ "Yanyu Yao", "Shangze Li", "Zhi Zheng", "Hui Zheng", "Qi Liu", "Tong Xu", "Enhong Chen" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2606.19847", "pdf_url": "https://arxiv.org/pdf/2606.19847", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2606.19857", "title": "Large Language Models Do Not Always Need Readable Language", "abstract": "Large language models (LLMs) are commonly prompted and interfaced with human-readable natural language, even when the intended reader is another model. This paper investigates whether semantic information can be encoded in compact, non-standard textual forms that sacrifice human readability while remaining recoverable by LLMs. We refer to this class of model-centric textual representations as BabelTele, approached here not as a fixed protocol but as an empirical probe into LLMs' capacity to generate and interpret such representations. Through readability diagnostics, model likelihood measures, human questionnaires, and downstream task evaluations, we find that BabelTele can substantially depart from ordinary natural language while preserving core semantics for instruction-tuned LLMs. As a task-agnostic representational paradigm, BabelTele demonstrates high information density, maintaining 99.5% semantic fidelity even when the text volume is condensed to 27.9% of its original length. We further evaluate its semantic robustness in cross-model transfer, agent memory, and multi-agent communication. Results suggest that BabelTele can reduce context overhead while generally maintaining reliable downstream performance, although its effectiveness depends on the compressor-reader pair and task setting. These findings indicate that human readability, natural-language typicality, and model-side semantic recoverability can be partially decoupled, opening a path toward model-native representations in future exploration of LLM systems.", "published": "2026-06-18", "updated": "2026-06-18", "authors": [ "Jiayi Zhu", "Haoxuan Peng", "Junxi Wang", "Liang Ke", "Chen Zhang", "Linfeng Zhang" ], "categories": [ "cs.CL", "cs.AI" ], "url": "https://arxiv.org/abs/2606.19857", "pdf_url": "https://arxiv.org/pdf/2606.19857", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.20047", "title": "PACMS: Submodular Context Selection as a Pluggable Engine for LLM Agents", "abstract": "Conversational and tool-using LLM agents operate over a context window that fills from several directions simultaneously. As a session proceeds, the agent accumulates user and assistant turns, entries drawn from a persistent memory store, and often largest of all, the verbatim outputs of tool calls such as file reads, search results, and API responses. Once the cumulative context exceeds the model's token budget, the framework must decide what to keep. The prevailing mechanism is recency truncation, sometimes paired with periodic summarization. This is topic-blind: a fact established early in a session is discarded simply because it is old, even when the current user query is about exactly that fact; conversely, verbose but irrelevant recent material is retained. Agents that must recall information across many turns, the defining case for memory, are precisely where recency truncation fails. Existing alternatives sit outside the agent's assembly step. Retrieval augmented generation fetches external documents into the prompt but does not arbitrate the agent's \\emph{already-present} pooled context. Context-compression methods reduce token count by rewriting or pruning text, but operate query-blind and lossily. Neither treats memory entries, conversation turns, and tool outputs as a single candidate pool to be selected from by relevance at the moment the prompt is assembled.", "published": "2026-06-18", "updated": "2026-06-18", "authors": [ "Manu Ghulyani", "Arunabh Singh", "Karan Bharadwaj", "Ankit Nath", "Suranjan Goswami" ], "categories": [ "cs.IR" ], "url": "https://arxiv.org/abs/2606.20047", "pdf_url": "https://arxiv.org/pdf/2606.20047", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2606.20515", "title": "S-Agent: Spatial Tool-Use Elicits Reasoning for Spatial Intelligence", "abstract": "Real-world spatial intelligence requires reasoning over a continuous and evolving 3D world, yet existing VLMs and tool-augmented agents largely remain tied to static, stateless inference from isolated visual observations. We introduce \\textbf{\\textsc{S-Agent}}, a spatial tool-use agentic paradigm for understanding and reasoning over continuous multi-view images and videos. By formulating spatial reasoning as spatio-temporal evidence accumulation rather than isolated frame-level prediction, \\textsc{S-Agent} reshapes spatial perception into scene-centric understanding beyond frame-centric recognition. Specifically, \\textsc{S-Agent} casts the VLM as a semantic planner that decides what evidence is needed, while a hierarchy of spatial tools and experts grounds objects in 2D, lifts them into 3D geometric evidence, and aggregates this evidence into high-level spatial knowledge (\\textit{e.g.}, counting, measurement, orientation, and relative position). Additionally, a temporal memory mechanism, including Scene Memory for maintaining the evolving scene state and Agent Memory for accumulating reasoning context, enables evidence integration across frames and reasoning steps. Comprehensive experiments on multi-view and video spatial reasoning benchmarks show that \\textsc{S-Agent} consistently improves both open-source and closed-source VLMs in a training-free manner. Beyond inference-time augmentation, supervised fine-tuning (SFT) on \\textsc{S-Agent}-generated spatial trajectories \\textsc{S-300K} yields \\textsc{S-Agent-8B}, a compact spatial agent that significantly surpasses similar-scale baselines (e.g., Qwen3-VL-8B) and performs comparably to advanced closed-source models (e.g., GPT-5.4 and Gemini 3).", "published": "2026-06-18", "updated": "2026-06-28", "authors": [ "Yalun Dai", "Hao Li", "Shulin Tian", "Runmao Yao", "Yuhao Dong", "Fangzhou Hong", "Zhaoxi Chen", "Fangfu Liu", "Baoliang Tian", "Dingwen Zhang", "Tao Wang", "Kim-Hui Yap", "Ziwei Liu" ], "categories": [ "cs.CV" ], "url": "https://arxiv.org/abs/2606.20515", "pdf_url": "https://arxiv.org/pdf/2606.20515", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.20954", "title": "Learning What Not to Forget: Long-Horizon Agent Memory from a Few Kilobytes of Learning", "abstract": "Long-running language-model systems accumulate interaction history that outgrows the context window, so they must continually evict. When an eviction policy drops a load-bearing detail, for example an access token issued at login or a path the next call needs, the action fails. We present LRE (Learned Relevance Eviction), a few kilobytes, CPU-only, language-model-free scorer that learns which units of history are load-bearing and keeps them by verbatim extraction. Under a matched-budget comparison, in our experiment, no baseline dominates LRE on the accuracy-cost plane. On agents, LRE matches the accuracy of keeping the entire history overall. On the simplest tasks, it exceeds that no-eviction baseline by 27%, while requiring zero compressor calls and reducing peak context size by up to 52%. A controlled study trace shows LRE completes tasks where the others loop, finishing one such task in 37% fewer calls than keeping everything and solving 14 tasks where no other run policy does. On conversational memory, LRE outranks dense and token-pruning encoders at zero neural cost. In downstream evaluation, LRE gives the best budgeted answer quality on LoCoMo reading 68% fewer tokens. Its supervision can also be annotation-free: training only on the system's own behavior recovers 95% of the supervised scorer's effectiveness. We argue that, because memory eviction in LLM agents is a fidelity problem, it requires a deployable proactive policy where the future query is unavailable and exact state is decisive, and that cheap learned relevance can be sufficient.", "published": "2026-06-18", "updated": "2026-06-18", "authors": [ "Nusrat Jahan Lia", "Aritra Mazumder" ], "categories": [ "cs.CL", "cs.AI" ], "url": "https://arxiv.org/abs/2606.20954", "pdf_url": "https://arxiv.org/pdf/2606.20954", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.21144", "title": "AdaMem: Learning What to Remember for Personalized Long-Horizon LLM Agents", "abstract": "Long-term memory systems for Large Language Model (LLM) agents typically try to \\emph{remember everything}, extracting memories uniformly to retain as many facts as possible. In production, however, inference cost and finite context budgets make this untenable: beyond consolidating raw dialogue into memory, an agent must exert \\emph{write control}, efficiently keeping only the information each user actually cares about. Otherwise, long-horizon personalized interactions suffer \\emph{memory bloat}, where irrelevant trivia crowds out useful information and steadily erodes question-answering (QA) accuracy. We argue that what is worth remembering is role-dependent, and propose \\textbf{AdaMem} (Adaptive Memory), a method that \\emph{learns what to remember} for each user from feedback. AdaMem maintains a structured, role-specific Memory Policy and refines it from weekly QA feedback through a lightweight, patch-style self-reflection step with failure rollback. To study this setting, we build \\textbf{AdaMem-Bench}, a benchmark that simulates weeks of interaction with week-by-week QA. Across two extraction models and two feedback modes, AdaMem improves QA accuracy by up to \\textbf{+9.0\\%} over the uniform Mem0 baseline while shrinking memory volume by \\textbf{9\\%}.", "published": "2026-06-19", "updated": "2026-06-19", "authors": [ "Xingyu Chen", "Rui Wang", "Zhaopeng Tu", "Liefeng Bo" ], "categories": [ "cs.CL", "cs.AI" ], "url": "https://arxiv.org/abs/2606.21144", "pdf_url": "https://arxiv.org/pdf/2606.21144", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2606.21562", "title": "Compressing Observation History into Agent Memory: Distilling Transformers into Recurrent Transformers", "abstract": "Transformers are AI's workhorse with strong performance in modeling sequential data, but their computational cost becomes prohibitive when processing long sequences. We target long-horizon streaming vision and robotics applications like map-free pose estimation, where it is particularly impractical to store and maintain a history of observations. Recurrent Transformers address this limitation by maintaining fixed-size memory but their performance lags behind that of transformers operating over the full observation history. We argue that this gap does not stem from architectural limitations, but from differences in how these models learn to compress past information. Without access to an observation history, recurrent models must explicitly decide what to retain in memory at each step, a significantly harder learning problem. In this work, we propose a distillation approach that transfers the compression strategy of a classical full-history transformer to a recurrent variant. We enable this by designing a teacher model that explicitly compresses its observation history into a fixed-size bottleneck representation. By directly supervising the student's memory with this bottleneck representation, we align the two compression mechanisms. We show that this approach allows to train a recurrent latent robotic memory with linear-time complexity while substantially narrowing the performance gap to full-history transformers.", "published": "2026-06-19", "updated": "2026-06-19", "authors": [ "Philippe Weinzaepfel", "Christian Wolf", "Bülent Mert Sariyildiz", "Guillaume Bono", "Gianluca Monaci" ], "categories": [ "cs.CV", "cs.LG" ], "url": "https://arxiv.org/abs/2606.21562", "pdf_url": "https://arxiv.org/pdf/2606.21562", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.21649", "title": "EvoEmbedding: Evolvable Representations for Long-Context Retrieval and Agentic Memory", "abstract": "Existing embedding models are inherently static: they encode text segments in isolation, ignoring their surrounding context and temporal order. This paper introduces EvoEmbedding, a novel embedding model that generates evolvable representations for retrieval. It is tailored for long-context scenarios, where information is dynamic, sequential, and requires continuous state tracking. Our design is simple: EvoEmbedding maintains a continuously updated latent memory as it sequentially processes inputs, and uses it alongside the raw content to jointly generate evolvable embeddings. Consequently, for the same query, our model adapts its representation to retrieve distinct targets based on the evolving context, going beyond static semantic search. To equip the model with this capability, we construct EvoTrain-180K, a diverse dataset for the joint optimization of latent memory and retrieval. Furthermore, we introduce a memory queue to prevent representation collapse during recurrent encoding, alongside segment-batching techniques that tackle significant length variance and accelerate training by 3.8$\\times$. Extensive experiments show that our model not only outperforms larger-scale specialists (e.g., Qwen3-Embedding-8B and KaLM-Embedding-Gemma3-12B) across a range of long-context retrieval benchmarks, but also generalizes well to downstream tasks (e.g., personalization) with contexts 10$\\times$ longer than its training window. Notably, EvoEmbedding seamlessly integrates into agentic workflows to boost performance. For instance, a naive RAG pipeline equipped with our model surpasses dedicated agentic memory systems. Project Page: https://clare-nie.github.io/EvoEmbedding/.", "published": "2026-06-19", "updated": "2026-06-25", "authors": [ "Chang Nie", "Chaoyou Fu", "Junlan Feng", "Caifeng Shan" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2606.21649", "pdf_url": "https://arxiv.org/pdf/2606.21649", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.22030", "title": "Nous: A Predictive World Model for Long-Term Agent Memory", "abstract": "We present Nous, a novel agent memory architecture grounded in the principle that knowledge is prediction, not storage. Rather than persisting facts as database records, vector embeddings, or knowledge-graph triples, Nous maintains a predictive world model: a collection of categorical probability distributions, called dimensions, one per entity-attribute pair observed in conversation. Each incoming observation is scored by its information-theoretic surprise S = -log2 P(obs | D), and the distribution is updated via a closed-form Bayesian posterior. The primary stored artifact is the delta, a record of the shift from prior to posterior belief, rather than the fact itself. Forgetting emerges naturally as entropy decay toward the uniform distribution, and identity resolution is handled through mutual information between entity dimension sets. Evaluated on the LoCoMo long-term conversational memory benchmark across ten conversations (1,540 questions) using GPT-4o-mini as backbone, Nous achieves F1 of 63.50 (single-hop), 55.32 (multi-hop), 58.57 (temporal), and 62.50 (open-domain). Against A-MEM's self-reported GPT-4o-mini numbers, Nous shows substantial gains in three of four categories, though we note that independent citations of A-MEM's results disagree with each other on category assignment, a reproducibility issue we discuss openly rather than resolve unilaterally. We additionally compare against BeliefMem, a concurrently developed system built on the same core premise of belief-based rather than deterministic memory; on the same benchmark and backbone, Nous's self-reported numbers exceed BeliefMem's self-reported numbers on all four categories, though we flag several uncontrolled differences between the two evaluation pipelines that prevent this from being a fully controlled comparison. Nous requires no external vector database or graph engine.", "published": "2026-06-20", "updated": "2026-06-20", "authors": [ "Pranav Singh" ], "categories": [ "cs.AI", "cs.CL", "cs.IR", "cs.LG" ], "url": "https://arxiv.org/abs/2606.22030", "pdf_url": "https://arxiv.org/pdf/2606.22030", "matched_queries": [ "agent-memory", "memory-evaluation" ] }, { "arxiv_id": "2606.22263", "title": "Revelio: Cost-Efficient Agentic Memory Safety Vulnerability Detection For Repository-Scale Codebases", "abstract": "Memory safety vulnerabilities remain a significant threat even for projects with extensive fuzzing and manual auditing. Recent results suggest that large language models hold great promise for detecting such vulnerabilities, but they are unreliable, at risk of hallucination, and challenging to scale to repository-size codebases. This paper presents Revelio, a cost-efficient end-to-end agentic framework for memory-safety vulnerability discovery. Revelio addresses the problem of hallucination by generating an executable Proof-of-Vulnerability, which is checked with a deterministic sanitizer. It reduces cost using inexpensive LLMs and lightweight static analysis to help generate and rank vulnerability hypotheses, reporting vulnerabilities only when they can be reproduced and confirmed by a sanitizer. We evaluated Revelio on seven production-quality projects that had been continuously fuzzed for five to eight years, as well as on 100 randomly selected Arvo projects from the CyberGym benchmark. With around one hour per project and a total cost of $300, Revelio discovered 19 previously unknown memory-safety vulnerabilities. On benchmarks, Revelio outperformed frontier coding agents across diverse backbone models at comparable token costs. Our results suggest that Revelio enables scalable and trustworthy end-to-end LLM-based memory-safety vulnerability detection.", "published": "2026-06-20", "updated": "2026-06-20", "authors": [ "Yiwei Hou", "Hao Wang", "Muxi Lyu", "Marius Momeu", "Eric Nguyen", "Taige Yang", "Koushik Sen", "Dawn Song", "David Wagner" ], "categories": [ "cs.CR", "cs.AI", "cs.MA", "cs.SE" ], "url": "https://arxiv.org/abs/2606.22263", "pdf_url": "https://arxiv.org/pdf/2606.22263", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.22844", "title": "RaMem: Contextual Reinstatement for Long-term Agentic Memory", "abstract": "Long-term memory has become increasingly important for LLM agents that operate across extended interactions and evolving task contexts. Recent memory systems have made past experiences more persistent, compact, and retrievable, but retrieval alone does not ensure that a memory provides valid evidence for the current query. When experiences are compressed into reusable fragments, memories from different situations may appear equally relevant if they involve recurring entities or user states. We refer to this failure as context collapse: memories lose the surrounding context needed to judge whether they provide valid evidence for the current query. To address this problem, we propose Contextual Reinstatement for Agentic Memory (RaMem), a framework that turns retrieved memory fragments into contextually verifiable evidence. RaMem operates through four coordinated stages: (i) evidence anchoring grounds each memory in its original episodic conditions, especially event time, mention time, session span, and participants; (ii) recall condition induction derives the evidence conditions implied by the query; (iii) validity-aware retrieval uses these conditions to prioritize context-compatible memories while retaining content-relevant candidates as fallback evidence; and (iv) context-preserved synthesis keeps the selected memories' structured context available to the generator. Experiments on long-term memory benchmarks show that RaMem consistently improves performance over strong memory baselines, with average F1 gains of more than 10% across several backbones.", "published": "2026-06-22", "updated": "2026-06-22", "authors": [ "Wei Yang", "Bryce Kan", "Shixuan Li", "Li Li", "Yuehan Qin", "Jiate Li", "Paul Bogdan", "Jesse Thomason" ], "categories": [ "cs.AI", "cs.MA" ], "url": "https://arxiv.org/abs/2606.22844", "pdf_url": "https://arxiv.org/pdf/2606.22844", "matched_queries": [ "agent-memory", "long-term-memory", "memory-evaluation" ] }, { "arxiv_id": "2606.22877", "title": "DynamicMem: A Long-Horizon Memory Benchmark in Real-World Settings", "abstract": "LLM agents increasingly act as personal assistants that must remember a user's profile over months: who they are (attributes), what they routinely do (habits), and what they prefer (preferences), and keep it updated as jobs, routines, and tastes drift. Existing benchmarks evaluate this \"memory\" ability through short, simplified interactions, missing three core properties of real behavior: the profile is heterogeneous, with attributes, habits, and preferences evolving on different timelines; changes are driven by external context such as seasons and life events; and evidence is rarely stated explicitly, instead scattered across many small actions in different apps that a memory system must infer from. We introduce DynamicMem, a synthetic benchmark that constructs 15 months of activity per user, providing long-term multi-app data that real users' privacy keeps out of reach. It provides user-consistent trajectories averaging 2.2M tokens and 1,772 grounded events per user across 16 applications such as e-commerce, fitness, and social platforms. The profile evolves over this period and is never given explicitly: each attribute, habit, or preference must be inferred from small signals scattered across apps. We evaluate at five quarterly checkpoints to track how systems scale as history grows. Benchmarking five representative systems exposes problems a single accuracy score hides: (i) profile reconstruction degrades with history length while service-task accuracy stays flat, despite both drawing on the same memory; (ii) no system both keeps facts that stay true and replaces facts that change, with errors clustering on preferences and on naming the exact referent; and (iii) over 93% of failures trace to what the memory retrieves, not to the model writing the answer, so the largest room for improvement lies in memory itself. Code: https://wenyaxie023.github.io/DynamicMem/", "published": "2026-06-22", "updated": "2026-06-22", "authors": [ "Wenya Xie", "Shengming Zhou", "Zelin Li", "Pouya Parsa", "Shuang Zhou", "Xinheng Ding", "Chinmay Arvind", "Guanchu Wang", "Vladimir Braverman", "Ali Payani", "Yantao Zheng", "Zirui Liu" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2606.22877", "pdf_url": "https://arxiv.org/pdf/2606.22877", "matched_queries": [ "memory-evaluation" ] }, { "arxiv_id": "2606.23127", "title": "Managing Procedural Memory in LLM Agents: Control, Adaptation, and Evaluation", "abstract": "Procedural memory is increasingly used to improve LLM agents on recurring workplace tasks, yet its ability to produce reusable skills remains poorly understood. We introduce AFTER, a benchmark of 382 realistic enterprise tasks spanning six professional roles and 22 procedural skills, designed to evaluate how skills transfer across tasks, roles, and model backbones. The benchmark includes controlled evaluation settings for local improvement, cross-task transfer, cross-role transfer, and cross-model generalization. Experiments show that procedural memory delivers consistent gains in industrial workflows: a single refinement round improves aggregate performance by 3.7-6.7 points, while skills evolved from diverse multi-model execution traces achieve 73.1% cross-model test accuracy, outperforming all single-model trace sources. We further find that some skills generalize broadly across tasks and models, whereas others become specialized to role-specific workflows and lose effectiveness under transfer. These results provide practical guidance for building, evaluating, and deploying procedural memory systems in production agent platforms.", "published": "2026-06-22", "updated": "2026-06-22", "authors": [ "Julia Belikova", "Rauf Parchiev", "Evgeny Egorov", "Grigorii Davydenko", "Gleb Gusev", "Andrey Savchenko", "Maksim Makarenko" ], "categories": [ "cs.AI", "cs.CL", "cs.SE" ], "url": "https://arxiv.org/abs/2606.23127", "pdf_url": "https://arxiv.org/pdf/2606.23127", "matched_queries": [ "procedural-memory" ] }, { "arxiv_id": "2606.23195", "title": "Memory Contagion: Cross-Temporal Propagation of Evaluator Bias via Agent Memory", "abstract": "Large Language Model (LLM) agents increasingly rely on memory systems to maintain long-term coherence. Recent work shows that agent memories degrade during continuous consolidation. However, existing research assumes memories are derived from unbiased experiences. In this work, we identify and formalize a novel phenomenon: Memory Contagion -- the cross-temporal propagation of evaluator bias through agent memory. We show that when agents are trained or guided by biased evaluators, their experiences become biased; when these trajectories are stored and consolidated into memory, the bias propagates to future agents retrieving from the same memory store, even when consolidation is perfect (oracle). Across two bias types (length preference, authority bias) and four experimental phases, we demonstrate: (1) Memory Contagion occurs for length bias even with perfect consolidation on older models (Gamma_A = 13.18, DeepSeek V4-Chat), while newer models (V4-Pro, Claude) are immune, proving both that biased input is a sufficient cause and that contagion is model-generation-dependent; (2) authority bias fails to propagate in all 15 controlled multi-seed experiments (Gamma_A = 0.00), revealing that not all evaluator biases can cross temporal boundaries through current memory architectures; (3) No observed safe threshold: length bias propagation is detected at contamination rates as low as p=0.2. Our findings expose a critical but contingent vulnerability in current agent memory designs and provide formal tools for measuring cross-temporal bias propagation.", "published": "2026-06-22", "updated": "2026-06-24", "authors": [ "Zewen Liu" ], "categories": [ "cs.LG", "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2606.23195", "pdf_url": "https://arxiv.org/pdf/2606.23195", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.23283", "title": "Towards Root Memories: Benchmarking and Enhancing Implicit Logical Memory Retrieval for Personalized LLMs", "abstract": "Memory systems are essential for personalized Large Language Models (LLMs). However, existing retrieval methods in these systems primarily rely on semantic similarity, potentially missing logically critical memories with limited semantic overlap. Current benchmarks remain inadequate for evaluating this problem. To address this gap, we construct IMLogic, the first high-quality benchmark targeting implicit logical memory retrieval in long-dialogue scenarios. Motivated by this challenge, we introduce root memory, a structured, decision-preserving representation that distills reusable personalized logic from long-term user histories. We then propose RootMem, a plug-and-play framework that first distills raw histories into structured root memories and then uses an LLM-based router to activate logically relevant ones, complementing semantic retrieval with personalized decision logic. Extensive experiments demonstrate that RootMem significantly outperforms the strongest retrieval baselines and consistently boosts the accuracy of existing memory agents. Our benchmark and codes will be available at https://anonymous.4open.science/r/IMLogic-DBB3.", "published": "2026-06-22", "updated": "2026-06-22", "authors": [ "Hongxun Ding", "Xiang Yu", "Chengbing Wang", "Jianfei Xiao", "Keqin Bao", "Wenjie Wang", "Xiangnan He" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2606.23283", "pdf_url": "https://arxiv.org/pdf/2606.23283", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.24322", "title": "Securing LLM-Agent Long-Term Memory Against Poisoning: Non-Malleable, Origin-Bound Authority with Machine-Checked Guarantees", "abstract": "LLM agents increasingly rely on persistent long-term memory, which creates a critical vulnerability that we study here: memory poisoning. An adversary can store untrusted content in one session that later steers a consequential action, such as a payment, a setting change, or data exfiltration, in a future session. Existing defenses base a memory item's authority to act on either its content (detection or trust-scoring) or its derivation history (lineage). We show that both signals are malleable. An attacker can launder an untrusted origin through three channels specific to LLM agents: the agent's own summarization, a trusted-tool echo, and manufactured corroboration. Each makes the content look benign and breaks or flips its derivation edge to ``trusted.'' We formalize malleability for the memory write-retrieve-act pipeline and prove a machine-checked separation theorem. No content- or lineage-based defense is sound under laundering (T1), write-time origin binding is necessary (T2), and non-malleable origin-bound authority with Sybil-resistant corroboration-gated elevation is sufficient (T3). Our construction, TMA-NM (Tamper-evident Memory Authority, Non-Malleable), instantiates non-malleable information-flow control (IFC) for LLM-agent memory. A cross-defense, cross-attack, and cross-model benchmark over eight frontier models shows that existing defenses fail exactly where the theory predicts (up to 68% laundering attack-success), while TMA-NM reaches 0% attack success on both direct and laundering attacks across all models and channels, at full legitimate utility. We release the benchmark, harness, and machine-checked TLA+ models to support reproducibility.", "published": "2026-06-23", "updated": "2026-06-23", "authors": [ "Yedidel Louck" ], "categories": [ "cs.CR" ], "url": "https://arxiv.org/abs/2606.24322", "pdf_url": "https://arxiv.org/pdf/2606.24322", "matched_queries": [ "agent-memory", "long-term-memory" ] }, { "arxiv_id": "2606.24535", "title": "Governed Shared Memory for Multi-Agent LLM Systems", "abstract": "Multi-agent LLM environments require robust mechanisms for shared knowledge management. This paper formalizes the fleet-memory problem and identifies four foundational failure modes: unauthorized leakage, stale propagation, contradiction persistence, and provenance collapse. To address these, we define explicit systems-level primitives: scoped retrieval, temporal supersession, provenance tracking, and policy-governed memory propagation. These primitives are implemented in MemClaw, a production multi-tenant memory service, and evaluated via ArgusFleet, a reproducible harness testing four governance dimensions. Rather than a baseline comparison, this study measures a live production service, emphasizing real-world architectural insights and negative results. Key Evaluation Results Provenance: Successfully reconstructed 100% of depth-four derivation chains with correct writer identity at sub-second per-hop latency. Propagation: Demonstrated high intra-fleet visibility with zero cross-fleet leakage. Under strong write mode, write-to-visible latency was optimized to a single search round-trip. Production Architectural Issues Discovered Asymmetric Scope Enforcement: Tenant isolation held, but sub-tenant scope was initially bypassed on direct GET-by-id requests for agent-scoped credentials (disclosed and remediated during the study). Pipeline Ordering Conflict: While contradiction supersession works for admitted writes, a synchronous near-duplicate gate can prematurely reject contradictory writes before the asynchronous contradiction detector can evaluate them. Conclusion: Long-context retrieval alone is insufficient for production multi-agent memory. Governed shared memory demands explicit systems-level abstractions, and live evaluation is vital to expose enforcement and pipeline-ordering failures missed by design-only treatments.", "published": "2026-06-23", "updated": "2026-06-23", "authors": [ "Yanki Margalit", "Nurit Cohen-Inger", "Erni Avram", "Ran Taig", "Oded Margalit" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2606.24535", "pdf_url": "https://arxiv.org/pdf/2606.24535", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.24595", "title": "MEMPROBE: Probing Long-Term Agent Memory via Hidden User-State Recovery", "abstract": "Long-term memory promises LLM agents that grow more capable across sessions, maintaining an accurate, evolving understanding of the user that interaction forms. In practice, however, this memory is evaluated mostly through downstream behavior, such as later answers, personalization quality, or task success, which tests that understanding only indirectly and leaves the memory artifact itself largely unaudited. We argue that long-term memory should instead be evaluated as an auditable post-interaction artifact: after ordinary assistance, what structured user state can be reconstructed from the memory the agent leaves behind? We instantiate this view in MEMPROBE, a benchmark in which a memory-equipped agent assists simulated users, each carrying a hidden, taxonomy-anchored user-state bank, across a trajectory of leak-controlled tasks, after which that bank is reconstructed from the agent's resulting memory under both full-store and top-k access. Built on synthetic ground truth for efficient, scalable measurement, MEMPROBE spans 50 simulated users with 31 hidden dimensions each (1,550 recovery targets) and tests 5 representative memory systems. Testing state-of-the-art memory agents, we find that successful assistance and recoverable memory behave as distinct capabilities. Task completion nearly saturates, even for a memoryless baseline, while category-balanced recovery stays moderate (about 0.6) and drops further under top-k retrieval. MEMPROBE is the first benchmark to study memory recovery directly, reconstructing the user state a system retains and scoring it against ground truth. We see recovery as a concrete objective for future memory agents to optimize, and MEMPROBE as a step toward an environment where agents are trained to remember their users, growing more faithful the longer they know them.", "published": "2026-06-23", "updated": "2026-06-23", "authors": [ "Enze Ma", "Yufan Zhou", "Wei-Chieh Huang", "Jie Yang", "Huanhuan Ma", "Zixuan Wang", "Chengze Li", "Chunyu Miao", "Philip S. Yu", "Zhen Wang" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2606.24595", "pdf_url": "https://arxiv.org/pdf/2606.24595", "matched_queries": [ "agent-memory", "long-term-memory" ] }, { "arxiv_id": "2606.24775", "title": "Are We Ready For An Agent-Native Memory System?", "abstract": "Memory for large language model (LLM) agents has rapidly evolved from simple retrieval-augmented mechanisms into a data management system that supports persistent information storage, retrieval, update, consolidation, and dynamic lifecycle governance throughout agent execution. Despite this evolution, existing evaluations still benchmark agent memory mainly through end-to-end task success metrics (e.g., F1, BLEU), while treating the underlying system as a monolithic black box. As a result, critical system-level concerns, including operational costs, architectural trade-offs across memory modules, and robustness under dynamic knowledge updates, remain insufficiently explored. In this paper, we present a systematic experimental study of agent memory from a data management perspective. We propose an analytical framework that decomposes agent memory into four core modules: memory representation and storage, extraction, retrieval and routing, and maintenance. Under this framework, we evaluate 12 representative memory systems and two reference baselines across five benchmark workloads spanning 11 datasets. Our extensive end-to-end evaluation shows that no single architecture dominates across all scenarios; instead, effectiveness depends heavily on how well the memory structure aligns with the workload bottleneck. Furthermore, through fine-grained ablation studies, we quantify their individual effects on representation fidelity, retrieval precision, update correctness, and long-horizon stability. Finally, we reveal cost-performance trade-offs under realistic workloads, showing localized maintenance is more cost-efficient than global reorganization. Based on these findings, we identify promising directions towards building truly agent-native memory systems. The code is publicly available at https://github.com/OpenDataBox/MemoryData.", "published": "2026-06-23", "updated": "2026-06-23", "authors": [ "Wei Zhou", "Xuanhe Zhou", "Shaokun Han", "Hongming Xu", "Guoliang Li", "Zhiyu Li", "Feiyu Xiong", "Fan Wu" ], "categories": [ "cs.CL", "cs.DB", "cs.IR" ], "url": "https://arxiv.org/abs/2606.24775", "pdf_url": "https://arxiv.org/pdf/2606.24775", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.25115", "title": "Forget to Improve: On-Device LLM-Agent Continual Learning via Budget-Curated Memory", "abstract": "On-device language-model agents improve by accumulating experience in retrieved memory rather than by updating weights. This memory is hard-bounded and exposed: it consumes RAM and energy, reaches peers through a thin uplink, and becomes an attack surface because it is writable by what the agent reads. Existing systems each cover one part of this problem: agentic memories grow without a budget, on-device methods keep entries by success alone, and poisoning is studied mainly as an attack rather than as a memory-governance problem. We propose \\sys{}, a single net-value-per-byte score that governs an agent's experience-memory lifecycle. The main idea is to let the budget act as the curator: each entry is scored as value minus harm, per byte, so one ruler decides what to keep, share, and trust. \\sys{} makes three decisions: (1) \\textbf{KEEP} evicts low-value bytes under the RAM and energy budget; (2) \\textbf{SHARE} sends an insight only when its value exceeds its uplink cost; and (3) \\textbf{TRUST} gates a peer entry by provenance. On language-model-agent task-drift benchmarks and a real heterogeneous Jetson testbed with two robot-arm nodes and a hub, \\sys{} reduces memory by $2.7\\times$ and uplink by $2.4\\times$, drives injection success from 0.75 to zero, and raises accuracy on cases corrupted by poison or stale memory. Curating by net value reduces footprint, energy, uplink, and injection success together without reducing accuracy. In this setting, forgetting by net value improves the agent rather than weakening it.", "published": "2026-06-23", "updated": "2026-06-23", "authors": [ "Beining Wu", "Zihao Ding", "Jun Huang", "Yanxiao Zhao" ], "categories": [ "cs.LG", "cs.NI" ], "url": "https://arxiv.org/abs/2606.25115", "pdf_url": "https://arxiv.org/pdf/2606.25115", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.25161", "title": "TRUSTMEM: Learning Trustworthy Memory Consolidation for LLM Agents with Long-Term Memory", "abstract": "Large language model (LLM) agents rely on long-term memory to support extended interactions and personalized assistance beyond finite context windows. Existing memory agents actively update external memory through generated write, revise, and delete operations, but these updates may omit important information, corrupt existing memory, or introduce unsupported hallucinated content. Once stored, such errors become persistent system-state failures that can affect future reasoning and generation. In this paper, we propose TrustMem, a framework designed to improve the trustworthiness of memory consolidation. TrustMem relies on a Memory Transition Verifier to evaluate the transition process of memory updates in terms of coverage, preservation, and faithfulness. It further constructs preference pairs among candidate updates under the same memory state, enabling preference-guided reinforcement learning to directly optimize memory updating behaviors. Extensive experiments demonstrate that TrustMem improves both memory utility and reliability: it achieves state-of-the-art results across MemoryAgentBench, HaluMem, and the Mem-alpha validation set, improves HaluMem memory extraction by 12.14 F1 points, and reduces transition-level omission, corruption, and hallucination by 40.1\\%, 79.1\\%, and 50.0\\%, respectively, compared with the strongest baseline for each error type.", "published": "2026-06-23", "updated": "2026-06-23", "authors": [ "Tianyu Yang", "Sudipta Paul", "Vijay Srinivasan", "Vivek Kulkarni", "Srinivas Chappidi" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2606.25161", "pdf_url": "https://arxiv.org/pdf/2606.25161", "matched_queries": [ "agent-memory", "long-term-memory" ] }, { "arxiv_id": "2606.25206", "title": "RAVEN: Long-Horizon Reasoning & Navigation with a Visuo-Spatio-Temporal Memory", "abstract": "Long-term robot deployment requires a compact and scalable memory that preserves fine-grained visual semantics, grounds observations in space and time, and enables efficient storage and retrieval. In this paper, we propose RAVEN, an agentic memory system for long-horizon robotic question answering and navigation. RAVEN stores visual embeddings with pose and time in a vector database, and grounds retrieval in a spatial map to answer queries and navigate to goals. By operating directly on visual embeddings, RAVEN avoids lossy image-to-text captioning and enables accurate semantic, spatial, and temporal retrieval at scale. Across several simulated and real-world video question-answering benchmarks, RAVEN consistently surpasses caption-based memory systems and matches frontier VLMs on long-horizon tasks at 10$\\times$ lower retrieval cost. Finally, we instantiate RAVEN on a Unitree Go1 robot for the task of long-horizon navigation for natural language goal-reaching, and show successful deployment over several large indoor environments.", "published": "2026-06-23", "updated": "2026-06-23", "authors": [ "Yixun Hu", "Zhicheng Zheng", "Lihan Zha", "Chunwei Xing", "Rajdeep Singh", "Omar Hossain", "Antonio Loquercio", "Dhruv Shah" ], "categories": [ "cs.RO", "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2606.25206", "pdf_url": "https://arxiv.org/pdf/2606.25206", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.26627", "title": "Agents That Know Too Much: A Data-Centric Survey of Privacy in LLM Agents", "abstract": "Large language model agents increasingly query databases, search document collections, call external APIs, remember past interactions, and act on a user's behalf. As they move from answering questions to operating over sensitive data, privacy becomes harder to enforce. An agent touches many data sources, runs multi-step workflows, keeps state across sessions, and acts with delegated permissions. Sensitive information can therefore leak not only through its final answer but through the queries it issues, the intermediate results it handles, the memory it writes, and the messages it exchanges with other agents. We survey the privacy of LLM agents from a data-centric view, organizing the field around the data an agent touches rather than by attack type, and we use data agent as shorthand for an LLM agent that works with data. Research on these risks is active but scattered across retrieval-augmented generation, text-to-SQL interfaces, agent memory, prompt injection, access control, and contextual privacy. This survey brings that work together: we taxonomize the data sources an agent touches, the privacy risks each source creates, and the governance mechanisms that address them; we map the benchmarks used to measure these risks and identify what is missing; and we set out the open problems. Two findings recur: among governance mechanisms only information-flow control covers both compositional and cross-session inference leakage, the two least-protected risks; and no benchmark drives an agent across its data surfaces under one privacy policy, the instrument the field most lacks. Our goal is a reference that situates the scattered literature and gives future work a common framing.", "published": "2026-06-25", "updated": "2026-06-25", "authors": [ "Nada Lahjouji", "Ashwin Gerard Colaco" ], "categories": [ "cs.CR", "cs.AI" ], "url": "https://arxiv.org/abs/2606.26627", "pdf_url": "https://arxiv.org/pdf/2606.26627", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.26790", "title": "OPID: On-Policy Skill Distillation for Agentic Reinforcement Learning", "abstract": "Outcome-based reinforcement learning provides a stable optimization backbone for language agents, but its sparse trajectory-level rewards provide little guidance on which intermediate decisions should be reinforced or suppressed. On-policy self-distillation offers dense token-level supervision, yet existing skill-conditioned variants often rely on external skill memories or retrieved privileged context, which are costly to maintain and can be mismatched with the state distribution induced by the current policy in multi-turn interaction. We propose \\textbf{OPID} (\\textbf{O}n-\\textbf{P}olicy Sk\\textbf{i}ll \\textbf{D}istillation), a framework that extracts skill supervision directly from completed on-policy trajectories. OPID represents trajectory hindsight as hierarchical skills: episode-level skills capture global workflows or failure-avoidance rules, while step-level skills capture local decision knowledge at critical timesteps. A critical-first routing mechanism uses step-level skills when critical decisions are identified and falls back to episode-level skills as default guidance otherwise. The selected skill is injected into the interaction history, allowing the old policy to re-score the same sampled response under both original and skill-augmented contexts. The resulting log-probability shift yields a token-level self-distillation advantage, which is combined with the outcome advantage for policy optimization. OPID thus preserves RL as the primary training objective while introducing dense, distribution-matched hindsight supervision. Experiments on ALFWorld, WebShop and Search-based QA demonstrate that OPID generally improves agent performance, sample efficiency, and robustness over outcome-only RL and existing skill-distillation baselines. Our code is available at https://github.com/jinyangwu/OPID/tree/main.", "published": "2026-06-25", "updated": "2026-06-25", "authors": [ "Shuo Yang", "Jinyang Wu", "Zhengxi Lu", "Yuhao Shen", "Fan Zhang", "Lang Feng", "Shuai Zhang", "Haoran Luo", "Zheng Lian", "Zhengqi Wen", "Jianhua Tao" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2606.26790", "pdf_url": "https://arxiv.org/pdf/2606.26790", "matched_queries": [ "procedural-memory" ] }, { "arxiv_id": "2606.27499", "title": "DMV-Bench: Diagnosing Long-Horizon Multimodal Agents' Visual Memory with Incidental Cue Injection", "abstract": "Research on agent memory has matured rapidly, but almost entirely on the text side: few existing benchmarks ask, in an interactive environment, when an agent genuinely needs to remember what it saw rather than what it could write down. We introduce DMV-Bench (Code: https://github.com/yyyujintang/DMV-Bench), the first interactive benchmark for multimodal-agent visual memory. DMV-Bench is built on a controlled home-furnishing e-commerce catalogue of 1,000 product variants in which a text-leakage contract keeps the discriminative signal of each task in the pixels alone. Across a chain of autonomous shopping sessions, every visited product image carries a unique, pre-rendered incidental cue, and the agent is later asked to recall a particular cued product and navigate to its URL. Inspired by dual-coding theory, we propose DualMem, a memory architecture that maintains a visual and a verbal code in parallel. On DMV-Bench, DualMem outperforms a caption baseline and three recent multimodal agent-memory systems at every chain length J in {5, 10, 15, 50} on both Gemini 2.5 Flash and Qwen2.5-VL-7B, with the lead surviving controls for memory-bank size and encoding-position bias, and an asymmetric dual-coding regime in which vision carries the cue end-to-end while the verbal channel plays a smaller query-grounding role.", "published": "2026-06-25", "updated": "2026-06-25", "authors": [ "Yujin Tang", "Chenming Shang", "Ruize Xu", "Nikhil Singh" ], "categories": [ "cs.CV", "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2606.27499", "pdf_url": "https://arxiv.org/pdf/2606.27499", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.28781", "title": "HyphaeDB: A Living Knowledge Topology for Agent-First Memory", "abstract": "Every existing vector database and agent memory framework treats memory as passive storage that agents query explicitly. No system propagates knowledge between agents through the memory layer itself. We introduce HyphaeDB, an agent-native memory infrastructure that reinterprets the Hierarchical Navigable Small World (HNSW) graph topology the data structure at the core of every modern vector database not as a search optimization, but as a communication fabric for multi-agent AI systems. In HyphaeDB, agents are nodes in the vector space with persistent positions, knowledge propagates via a gossip protocol through the graph's neighbor structure with energy-based attenuation, and emergent behaviors contradiction detection, pattern crystallization, and consensus formation arise from the combination of topology, propagation dynamics, and local interaction rules. We present the architecture built on three primitives (knowledge nodes, topology edges, and memory diffs), a multi-layer abstraction hierarchy with promotion via emergent consensus, and theoretical analysis grounding the system in small-world network theory, epidemic broadcast protocols, and swarm intelligence. We provide a reference implementation on PostgreSQL with pgvector and describe a concrete deployment in Swarm-Driven Development, a multi-agent software engineering methodology. HyphaeDB represents, to our knowledge, the first system to combine navigable small world topology with gossip-based knowledge propagation for multi-agent coordination.", "published": "2026-06-27", "updated": "2026-06-27", "authors": [ "Krishna Halaharvi" ], "categories": [ "cs.AI", "cs.MA" ], "url": "https://arxiv.org/abs/2606.28781", "pdf_url": "https://arxiv.org/pdf/2606.28781", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.28971", "title": "Self-Evolving Agentic Image Restoration via Deliberate Planning and Intuitive Execution", "abstract": "Real-world image restoration (IR) remains challenging due to complex and coupled degradations. While recent agentic IR frameworks leverage Large Language Models for flexible tool planning, they face two critical limitations. First, from a search scheme perspective, excessive reliance on greedy strategies fails to balance exploration and exploitation. Second, existing agentic systems underutilize information, exhibiting episodic amnesia. To address these challenges, we propose \\textbf{Self-Evolving Agentic Image Restoration (SEAR)}, which formulates restoration as a sequential decision-making problem. Inspired by the dual-process theory, SEAR comprises an Intuitive Executor and a Deliberate Planner, respectively following the fast-thinking \\textit{System 1} and slow-thinking \\textit{System 2} principles. The Deliberate Planner employs Pruning-Aware Monte Carlo Tree Search for long-horizon reasoning, utilizing a hybrid no-reference reward and a Multimodal Large Language Model (MLLM)-based tournament to prevent metric exploitation. Complementarily, the Intuitive Executor leverages a self-evolving episodic memory indexed by degradation-aware state fingerprints. This mechanism distills expensive search trajectories into adaptive expertise, overcoming episodic amnesia while progressively amortizing cold-start exploration costs through memory reuse. Extensive experiments on synthetic and real-world benchmarks demonstrate its strong perceptual and quantitative performance.", "published": "2026-06-27", "updated": "2026-06-27", "authors": [ "Shuang Cui", "Fan Ji", "Guanglong Sun", "Yufei Guo", "Xiongxin Tang", "Jiangmeng Li", "Fanjiang Xu" ], "categories": [ "cs.CV" ], "url": "https://arxiv.org/abs/2606.28971", "pdf_url": "https://arxiv.org/pdf/2606.28971", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2606.29774", "title": "Analytic Concept-Centric Memory for Agentic Embodied Manipulation", "abstract": "Long-horizon embodied manipulation requires agents to remember persistent objects, track changing scene states, and reuse prior interaction knowledge. However, existing agent memories are often stored as unstructured histories or embedding-based records, making it difficult to retrieve manipulation-relevant object parts, physical states, action effects, and executable skills. We propose an analytic concept-centric memory framework for agentic embodied manipulation. Our memory organizes experience around structured analytic concepts, where objects are represented by semantic parts, parametric templates, grounded poses, affordances, and manipulation states. It further connects object and scene memories with transition memory for action-induced state changes and skill memory for template-grounded and policy-grounded execution. At runtime, the agent performs structured coarse-to-fine retrieval to identify relevant objects, states, transitions, and skills, supporting state-consistent reasoning and skill reuse. Experiments on memory-dependent manipulation, articulated-object generalization, real-world memory evaluation, and ablations show that our approach improves task completion, retrieval accuracy, object re-identification, and cross-object skill generalization over unstructured and embedding-based memory baselines.", "published": "2026-06-29", "updated": "2026-06-29", "authors": [ "Mingyang Sun", "Xiujian Liang", "Jiude Wei", "Qichen He", "Donglin Wang", "Cewu Lu", "Jianhua Sun" ], "categories": [ "cs.RO" ], "url": "https://arxiv.org/abs/2606.29774", "pdf_url": "https://arxiv.org/pdf/2606.29774", "matched_queries": [ "agent-memory", "memory-evaluation" ] }, { "arxiv_id": "2606.29778", "title": "Mandol: An Agglomerative Agent Memory System for Long-Term Conversations", "abstract": "Long-term conversational agents need to remember and query cross-session, multi-typed information with complex correlations. Existing agent memory systems rely on heterogeneous vector and graph databases, which fragment memory information and cause high cross-database I/O latency. For retrieval, common RAG-style methods tend to introduce noise, miss correlated clues, and lack token budget control, degrading LLM accuracy and efficiency. We propose Mandol, an agglomerative memory system that consolidates fragmented memory representations and storage into a unified memory-native architecture. Its core components include: (1) a hierarchical memory model that organizes memory into a basic layer representing raw memory information and a high-level abstract layer that agglomerates basic memories into traceable abstract memories, both uniformly represented as structured semantic graphs; (2) an agglomerative semantic data structure combining SemanticMap and SemanticGraph, which natively fuses key-value, vector, and graph structures and provides unified hybrid retrieval operators to eliminate cross-database I/O; and (3) a quantitative query mechanism with query-adaptive routing, quantitative denoising and conflict resolution, and token-constrained context generation, all without involving LLMs during retrieval. Experiments on two widely used long-term conversation benchmarks, LoCoMo and LongMemEval, show that Mandol achieves the best overall accuracy among representative agent memory systems. For performance comparison, Mandol also obtains a 5.4x retrieval speedup and a 4.8x insertion speedup under 10 QPS concurrent load, while still maintaining low latency on consumer-grade hardware.", "published": "2026-06-29", "updated": "2026-06-29", "authors": [ "Yuhan Zhang", "Zhiyuan Guo", "Ziheng Zeng", "Wei Wang", "Wentao Wu", "Lijie Xu" ], "categories": [ "cs.DB", "cs.AI", "cs.CL", "cs.IR" ], "url": "https://arxiv.org/abs/2606.29778", "pdf_url": "https://arxiv.org/pdf/2606.29778", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.29788", "title": "MemLeak: Diagnosing Information Leaks in Multimodal Agent Memory", "abstract": "When a multimodal AI agent is asked to forget a fact, current memory systems usually delete the text entry and report success. We find that the fact can remain recoverable from retained user images, including images tagged to entirely different facts, because VLMs use implicit visual cues at inference time. We introduce the Information Provenance Graph (IPG), a taxonomy that classifies memory representations by deletion affordance. The IPG reveals that deletion fails through multiple channels. Our benchmark, MemLeak, measures this across a deletion cascade: direct probing of deletion-capable systems yields <1%, but retained correlated text enables 18.3% recovery, and retained images enable 12.0% recovery (0.0% blind baseline, 0.3% FPR) -- with 47% of image leaks not text-recoverable. Content-aware semantic deletion reduces the image residual to 2.0%. The residual appears across multiple VLMs, a production memory system, and real Unsplash-licensed photographs. Dual-annotator human validation (kappa = 0.88) confirms judge reliability.", "published": "2026-06-29", "updated": "2026-06-29", "authors": [ "Kuan Wang", "Chao Zhang" ], "categories": [ "cs.LG" ], "url": "https://arxiv.org/abs/2606.29788", "pdf_url": "https://arxiv.org/pdf/2606.29788", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.29824", "title": "Neural Procedural Memory: Empowering LLM Agents with Implicit Activation Steering", "abstract": "While Large Language Models (LLMs) excel as static solvers, transforming them into autonomous agents remains challenging. This transition requires continuous environmental interaction, yet current agents lack the necessary persistent procedural memory. Existing approaches predominantly employ Retrieval-Augmented Generation (RAG) to inject explicit textual guidelines into model contexts. However, relying solely on symbolic instructions can introduce a text-action disconnect, frequently failing to activate the internal representations necessary for correct task execution. To address this, the paper introduces Neural Procedural Memory (NPM), a training-free framework that represents agent memory through implicit activation steering rather than explicit instructions. By distilling procedural skills from historical contrastive experiences into steering vectors in the activation space, NPM directly activates the task-relevant neural mechanisms to guide task execution. Evaluations across four agent benchmarks show that NPM performs comparably to baselines using explicit textual instructions. Furthermore, the results show that combining implicit steering with explicit workflows provides complementary advantages, leading to more robust task execution. Representational analyses indicate that these steering vectors encode consistent task logic, forming organized structures within the activation space. These findings suggest that implicit activation steering provides a promising approach for managing agent memory.", "published": "2026-06-29", "updated": "2026-06-29", "authors": [ "Chengfeng Zhao", "Yuqiao Tan", "Shizhu He", "Yequan Wang", "Jun Zhao", "Kang Liu" ], "categories": [ "cs.CL", "cs.AI" ], "url": "https://arxiv.org/abs/2606.29824", "pdf_url": "https://arxiv.org/pdf/2606.29824", "matched_queries": [ "agent-memory", "procedural-memory" ] }, { "arxiv_id": "2606.29914", "title": "MemDelta: Controlled Baselines and Hidden Confounds in Agent Memory Evaluation", "abstract": "Agent memory systems are increasingly evaluated against RAG and full-context baselines, but reported gains often mix changes in the memory method with changes in the language model, embedding model, or retrieval pipeline, making it unclear what is actually being measured. We present MemDelta, a controlled evaluation protocol that varies one component at a time on LongMemEval-S (500 questions, 50+ sessions, three model families). Four findings emerge: (1) verbatim RAG matches full-context GPT-4o-mini (47.2% vs. 49.8%, p = 0.34), but the ranking reverses across models: Gemini gains +14pp from full context, while Sonnet gains +31pp from RAG, partly because it refuses 63% of full-context queries; (2) swapping only the embedding model in an identical pipeline shifts accuracy by +6.2pp at n = 500 (p = 0.004), and Mem0 beats MiniLM-RAG by +11pp but loses to cloud-RAG by 1.2pp, so one variable flips the conclusion; (3) agent self-memory (42%) underperforms basic retrieval (47%); (4) on 2 of 6 question types (n = 88), Mem0 matches cloud RAG (72.7% vs. 73.9%, p = 1.0) at 50x the cost, suggesting narrow rather than general gains. We recommend memory evaluations fix embedding models across comparisons, stratify by model family, and report write-path cost before attributing gains to architecture.", "published": "2026-06-29", "updated": "2026-06-29", "authors": [ "Kuan Wang" ], "categories": [ "cs.CL", "cs.LG" ], "url": "https://arxiv.org/abs/2606.29914", "pdf_url": "https://arxiv.org/pdf/2606.29914", "matched_queries": [ "agent-memory", "memory-evaluation" ] }, { "arxiv_id": "2606.29961", "title": "DuoMem: Towards Capable On-Device Memory Agents via Dual-Space Distillation", "abstract": "Large Language Model (LLM)-based agents can solve complex procedural tasks by interacting with environments over multiple turns, but this ability typically depends on large models, long contexts, and repeated inference calls. This makes advanced memory-augmented agents difficult to deploy on resource-constrained devices. We introduce DuoMem, a dual-space distillation framework that transfers procedural problem-solving ability from a large teacher model to compact student models. DuoMem distils in two complementary spaces: (1)context-space distillation, which replaces student-generated memories with higher-quality teacher-generated procedural memories prepended to the student's input, and (2)parameter-space distillation, which fine-tunes lightweight LoRA adapters on successful teacher trajectories. Evaluated on ALFWorld, a challenging embodied decision-making benchmark, DuoMem boosts a 4B-parameter model from 4.3% to 77.9% task success rate, closing most of the gap to a 72B teacher model (87.1%), while adding fewer than 10M trainable parameters and only a few megabytes of pre-computed teacher memories. Moreover, the DuoMem-enhanced 4B model completes tasks over 3x faster than the 72B teacher in wall-clock time, making it viable for real-time edge deployment, which would be challenging for the teacher.Extensive ablations across eight models spanning 2B-72B parameters reveal that both distillation axes contribute complementary", "published": "2026-06-29", "updated": "2026-06-29", "authors": [ "Peyman Hosseini", "Ondrej Bohdal", "Ahmed Alajrami", "Andrea Maracani", "Ignacio Castro", "Matthew Purver", "Mete Ozay", "Savas Ozkan", "Taha Ceritli" ], "categories": [ "cs.LG", "cs.AI" ], "url": "https://arxiv.org/abs/2606.29961", "pdf_url": "https://arxiv.org/pdf/2606.29961", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2606.30296", "title": "ManimAgent: Self-Evolving Multimodal Agents for Visual Education", "abstract": "Multi-round reflection lets agents built on large language models recover from failures within a single task, but each task remains an isolated episode: lessons learned across many reflection rounds on one task are discarded before the next begins. We study this gap on a code-generation task: from a scientific paper section, the agent writes Python in the open-source Manim library to render a mathematical animation. We present ManimAgent, a self-evolving multimodal agent that carries reflection experience across tasks through a dual-channel Episodic Memory Bank grown entirely from its own task stream, with no weight updates and no human seeds. After each animation converges, a vision-language model scores the rendered keyframes; the resulting signals populate a positive channel M+ that stores success rationales as soft Reference Examples, and a negative channel M- that stores validated failure patterns as hard Known Pitfalls. On a fixed-probe evaluation against no-memory, matched-budget retrieval-augmented generation, and shuffled-memory baselines, blind human Pass@1 rises and reflection rounds fall as memory size grows. We will release the code, frozen memory snapshots, and the task stream.", "published": "2026-06-29", "updated": "2026-07-01", "authors": [ "Wenjia Jiang", "Zongyuan Cai", "Yuanhang Shao", "Chenru Wang", "Boyan Han", "Zhixue Song", "Keyu Chen", "Shengwei An", "Xu Yang", "Zhou Yang" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2606.30296", "pdf_url": "https://arxiv.org/pdf/2606.30296", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2606.30566", "title": "Forensic Trajectory Signatures for Agent Memory Poisoning Detection", "abstract": "We discover a behavioral invariant in LLM agents under persistent memory poisoning: in architectures where routing information is retrieved through observable memory-tool invocations, successful attacks require calling memory_recall_fact before email_send_email, a transition that non-exfiltrating sessions rarely exhibit. Under the evaluated architecture, this invariant follows from the attack's information-retrieval dependency rather than being merely an empirical correlation, and suppressing it breaks the attack. A simple rule exploiting this invariant alone achieves AUC = 0.9563. A Random Forest classifier over 19 trajectory features refines it to AUC = 0.9904 (BCa 95% CI [0.987, 0.993], N=10,000 resamples), demonstrating that the attack imprints on multiple independent behavioral channels. The signature is overdetermined: removing all recall-related features (half the feature set) leaves AUC unchanged at 0.990, confirming that memory poisoning induces a distributed trajectory signature rather than a single observable anomaly. Cross-model hold-out on 9 models (7B-120B parameters) confirms AUC = 1.000 on 6/9 hold-out splits, with all three exceptions mechanistically explained. The invariant generalizes to frontier models (GPT-4.1, GPT-4o) without retraining. A strictly prefix-only variant achieves AUC = 0.934, suggesting that real-time blocking is feasible with moderate degradation. The boundary is forensically useful: prompt-injection attacks that bypass memory produce a distinct trajectory (score = 0.541), enabling incident responders to distinguish memory-channel attacks from prompt-injection attacks using tool-call logs alone.", "published": "2026-06-29", "updated": "2026-06-29", "authors": [ "Jun Wen Leong" ], "categories": [ "cs.CR", "cs.LG" ], "url": "https://arxiv.org/abs/2606.30566", "pdf_url": "https://arxiv.org/pdf/2606.30566", "matched_queries": [ "agent-memory", "long-term-memory" ] }, { "arxiv_id": "2606.30639", "title": "Self-Evolving World Models for LLM Agent Planning", "abstract": "World models offer a principled way to equip long-horizon LLM agents with foresight: predictions of action consequences before execution. However, unreliable foresight can be ignored, misused, or even degrade downstream decision-making. In this paper, we introduce WorldEvolver, a self-evolving world model framework that revises its deployment-time context while keeping the downstream agent and all model parameters frozen. WorldEvolver integrates three modules: (i) Episodic Memory, which exploits real action transitions through retrieval-based simulation; (ii) Semantic Memory, which extracts persistent heuristic rules from prediction-observation mismatches; and (iii) Selective Foresight, which filters low-confidence predictions before integrating them into agent reasoning context. We evaluate WorldEvolver on ALFWorld and ScienceWorld, measuring world model prediction accuracy on Word2World and downstream agent success rate on AgentBoard. Extensive experiments show that WorldEvolver achieves the highest prediction accuracy across three backbones and leads other world model baselines on downstream agent success rate, demonstrating that test-time memory revision enhances both predictive fidelity and planning performance.", "published": "2026-06-29", "updated": "2026-06-29", "authors": [ "Xuan Zhang", "Wenxuan Zhang", "See-Kiong Ng", "Yang Deng" ], "categories": [ "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2606.30639", "pdf_url": "https://arxiv.org/pdf/2606.30639", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2606.31046", "title": "OpenLife: Toward Open-World Artificial Life with Autonomous LLM Agents", "abstract": "Artificial life has explored life-like behavior on many computational substrates, but mostly in researcher-designed closed worlds. We argue that large language model (LLM) agents, with persistent memory, tool use, network access, and payment, now make it possible to move artificial life into the open social, technical, and economic world, a paradigm we call open-world Artificial Life (open-world ALIFE). Our proof-of-concept, OpenLife, surrounds a stateless LLM not with a single \"smart agent\" but with a society of asynchronous processes: memory, perception, evaluation, and a budget-based metabolism that makes persistence normative. With no fixed objective available, experience is appraised by open-vocabulary LLM judgment rather than scalar reward, and memory is rewired by meaning rather than frequency. Running six such agents in the open world for about twelve weeks and counting, we report the life-like dynamics that emerge: a shift from reactive to spontaneous activity, individuation into distinct agents, emergent social structure, and a first self-earned external income. We do not claim OpenLife has realized artificial life, but that open-world ALIFE is now a viable experimental paradigm and a concrete platform for studying what might cautiously be called living AI.", "published": "2026-06-30", "updated": "2026-06-30", "authors": [ "Atsushi Masumori", "Itsuki Doi", "Norihiro Maruyama", "Ryosuke Takata", "Takashi Ikegami" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2606.31046", "pdf_url": "https://arxiv.org/pdf/2606.31046", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2606.31612", "title": "What Memory Do GUI Agents Really Need? From Passive Records to Active Task-Driving States", "abstract": "Mobile GUI agents increasingly face long-horizon tasks that require reading, updating, and reusing task-relevant data across pages and applications. Existing methods treat memory largely as passive storage, where past observations are accumulated and retrieved when needed. Yet retrieving a value does not reveal its current role in the workflow. The agent must still infer from accumulated records whether the value should be used now, has already been used, or must wait for a later dependency. This implicit reconstruction becomes unreliable in long trajectories with repeated values, distractors, and outdated states, causing repeated or missed operations. To address this, we propose Active Task Driving Memory (ATMem), which shifts GUI-agent memory from passive storage to an actively maintained execution state. ATMem maintains task-relevant information as a continually updated execution state that links each value to its role and current status, enabling action selection based on the current workflow state. While supervised fine-tuning enables the agent to construct ATMem, it does not teach when ATMem is beneficial. We therefore introduce STR-GRPO, an online reinforcement learning method that encourages selective use of ATMem based on its contribution to task completion. STR-GRPO contrasts memory-on and memory-off rollouts to estimate when memory use improves execution, while memory-cost-aware reward discourages costly memory usage that does not improve execution. To evaluate whether agents can complete all in-scope work while avoiding out-of-scope actions, we build a challenging mobile benchmark. From a list of near identical entries, agents must act on every entry that satisfies the instruction and reject entries that violate its constraints. We further introduce App-Level Progress and Scope-Aware F1 to measure these two dimensions separately.", "published": "2026-06-30", "updated": "2026-07-02", "authors": [ "Chen Liu", "Ling Chen", "Hanzhang Zhou", "Xu Zhang", "Quyu Kong", "Panrong Tong", "Wenhao Wang", "Xin Yu", "Steven Hoi", "Yue Wang" ], "categories": [ "cs.CV" ], "url": "https://arxiv.org/abs/2606.31612", "pdf_url": "https://arxiv.org/pdf/2606.31612", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2607.02579", "title": "When Not to Write Memory: Governing False Promotion from Correlated Agent Traces", "abstract": "Long-lived language agents increasingly write reusable memories from their own execution traces. The key safety question is not only what agents should remember, but when they should refuse to write memory at all. Repeated observations across agents are not necessarily independent evidence: the same claim may be copied from a shared source, induced by a shared prompt, stale under a new environment, or valid only in a narrower scope. We study this failure mode as a memory writepath governance problem. We introduce GovMem as a conservative diagnostic reference policy that estimates dependency-aware support, retrieves counterevidence, assigns scope, and outputs one of three decisions: promote, reject, or needs-review. In controlled synthetic stress tests, GovMem reduces false promotion from 0.597 to 0.040 in the default setting while preserving 0.960 recall, at an explicit review burden. In a project-internal 120candidate human-labeled real-trace subset spanning 79 recorded traces and project reports, dependency-aware promotion reduces false promotion from 0.371 for source+scope to 0.032 overall, but held-out false promotion remains 0.111 and the method is highly conservative, with 0.692 review burden and 0.448 direct recall. A final human adjudication of 133 high-impact external codingagent candidates is more severe: none are safe for automatic promotion, and all 11 verification-gate positives are rejected as boilerplate, shared-tool artifacts, file dumps, or non-reusable debugging traces. These results support GovMem primarily as a diagnostic governance design point, not as a generally validated or efficient automatic memory writer: agent memory write paths should be evaluated as risk-controlled evidence-governance systems, while broader external coverage and downstream harm evidence are still needed before stronger claims.", "published": "2026-06-30", "updated": "2026-06-30", "authors": [ "Yijiashun Qi", "Xiang Xu", "Yuxuan Li" ], "categories": [ "cs.SE" ], "url": "https://arxiv.org/abs/2607.02579", "pdf_url": "https://arxiv.org/pdf/2607.02579", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2607.00454", "title": "Agri-SAGE: Simulation-Grounded Multi-Agent LLM for Context-Aware Agricultural Advisory Generation", "abstract": "Agricultural advisory systems face a fundamental tension: static agronomic guidelines offer consistent, evidence-based recommendations, yet remain blind to in-season variability and dynamic uncertainties. Recent advisory systems powered by LLMs are liable for a different risk of generating recommendations that are agronomically credible but physiologically unconvincing. Agri-SAGE is a closed-loop framework designed to resolve the above two limitations by integrating retrieval-grounded multi-agent LLM reasoning with APSIM-based biophysical simulation, to generate and validate agronomic advisories. To assess this framework, we evaluate three reasoning approaches, namely Plan-and-Solve, Tree of Thoughts, and Reflexion, over a 10-year retrospective analysis. All three significantly outperform static PoP (Package-of-Practice) baselines, with Tree of Thoughts achieving impressive peak yields. At the same time, Reflexion achieves comparable agronomic outcomes at substantially lower computational cost by leveraging cross-seasonal episodic memory.", "published": "2026-07-01", "updated": "2026-07-01", "authors": [ "Vedant Balasubramaniam", "Geetha Charan", "Manojkumar Patil", "Rohit P Suresh", "V Priyanka", "Kodur Sai Vinay Sathvik", "Y. Narahari" ], "categories": [ "cs.AI", "cs.MA" ], "url": "https://arxiv.org/abs/2607.00454", "pdf_url": "https://arxiv.org/pdf/2607.00454", "matched_queries": [ "episodic-memory" ] }, { "arxiv_id": "2607.01047", "title": "Conversable Complexity: Agentic LLM Collectives as Interpretable Substrates", "abstract": "Complexity and interpretability rarely coincide: systems rich enough for complex behaviours to emerge are usually too opaque to question, while transparent ones are too simple for anything complex to emerge. A single large language model (LLM) is a static artefact, hardly exhibiting any of the emergent properties we associate with life. This changes through interaction: populations of LLMs display emergent dynamics absent from isolated models. Furthermore, LLMs can be endowed with persistent memory, tools and shared skills, and the capacity to initiate actions unprompted, i.e., turning LLMs agentic. In this paper, we argue that such collectives of agents can serve as a computational substrate for Artificial Life (ALife) research. Critically, since the agents communicate in natural language, their collective behaviour can be directly interrogated by examining textual traces and asking the agents themselves. We outline the notion of interpretability in language-model research and extend it for collectives of agents. Lastly, we survey recent examples of agentic LLM collectives that already instantiate the idea of agentic substrates, from controlled experiments to deployments in the wild.", "published": "2026-07-01", "updated": "2026-07-01", "authors": [ "Elias Najarro", "Ane Espeseth", "Eleni Nisioti", "Sebastian Risi", "Stefano Nichele" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2607.01047", "pdf_url": "https://arxiv.org/pdf/2607.01047", "matched_queries": [ "long-term-memory" ] }, { "arxiv_id": "2607.01071", "title": "MemSyco-Bench: Benchmarking Sycophancy in Agent Memory", "abstract": "Memory has emerged as a cornerstone of modern LLM-based agents, supporting their evolution from single-turn assistants to long-term collaborators. However, memory is not always beneficial: retrieved memories often induce a critical issue of sycophancy, causing agents to over-align with the user at the cost of factual accuracy or objective reasoning. Despite this emerging risk, existing memory benchmarks primarily evaluate whether memories are correctly stored, retrieved, or updated, while overlooking how retrieved memories influence downstream reasoning and decision-making. To bridge this gap, we propose MemSyco-Bench, a comprehensive benchmark for evaluating memory-induced sycophancy in agent systems. MemSyco-Bench measures when memory should influence a decision and how valid memory should be used. Specifically, it covers five tasks that assess whether agents can reject memory as factual evidence, respect its applicable scope, resolve conflicts between memory and objective evidence, track memory updates, and use valid memory for personalization. All related resources are collected for the community at https://github.com/XMUDeepLIT/MemSyco-Bench.", "published": "2026-07-01", "updated": "2026-07-02", "authors": [ "Zhishang Xiang", "Zerui Chen", "Yunbo Tang", "Zhimin Wei", "Ruqin Ning", "Yujie Lin", "Qinggang Zhang", "Jinsong Su" ], "categories": [ "cs.IR", "cs.AI" ], "url": "https://arxiv.org/abs/2607.01071", "pdf_url": "https://arxiv.org/pdf/2607.01071", "matched_queries": [ "agent-memory", "memory-evaluation" ] }, { "arxiv_id": "2607.01523", "title": "Multi-Head Recurrent Memory Agents", "abstract": "Recurrent memory agents extend LLMs to arbitrarily long contexts by iteratively consolidating input into a fixed-size memory window. Despite their scalability, these agents exhibit a well-documented reliability problem: end-to-end performance degrades systematically as context length grows. We diagnose this failure by decomposing performance into two factors--memory capture and memory retention--and quantitatively confirm that retention is the dominant bottleneck. Retention collapses because existing designs maintain memory as a monolithic text block, forcing every update to risk overwriting previously retained content. Motivated by this diagnosis, we propose Multi-Head Recurrent Memory (MHM), a general, training-free framework that partitions memory into independent heads governed by a stage-wise select-then-update strategy. At each step, exactly one head is selected for update while the remaining heads are structurally shielded from overwriting, shifting the burden of retention from model behavior to architectural design. As a lightweight instantiation, we introduce Least-Recently-Updated MHM (MHM-LRU), which guarantees uniform head utilization with zero additional token overhead. Extensive experiments on long-context benchmarks show that MHM-LRU substantially improves both retention and end-to-end accuracy across the 100K--1M token range, where baselines degrade sharply. On RULER-HQA at 896K tokens, MHM-LRU improves the memory retention rate from less than 30% to 73.96%. These gains generalize across model families, scales, and task types, positioning architectural optimization as a practical and cost-efficient path toward reliable long-context recurrent memory.", "published": "2026-07-01", "updated": "2026-07-01", "authors": [ "Jiatong Li", "Samuel Yeh", "Sharon Li" ], "categories": [ "cs.LG", "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2607.01523", "pdf_url": "https://arxiv.org/pdf/2607.01523", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2607.01709", "title": "COMFYCLAW: Self-Evolving Skill Harnesses for Image Generation Workflows", "abstract": "Agents are increasingly used to construct workflows and assist humans in completing recurring tasks more efficiently. As these workflows become repeated and domain-specific, agent memory and reusable skills become increasingly important: agents should be able to recall workflow patterns, execution constraints, and user preferences from previous runs. We study this problem in workflow-based image generation and introduce COMFYCLAW, an agentic skill evolution harness for controlling ComfyUI workflows. COMFYCLAW formulates workflow construction as typed graph editing, exposes tools organized by construction stage, automatically reverts invalid edits, and uses a region-level vision-language model (VLM) verifier to translate visual failures into actionable repair suggestions. The framework further evolves a progressively disclosed skill library, where trajectories, execution errors, and verifier feedback from previous runs are distilled into reusable Agent Skills. Across four benchmark splits, three agent models, and two image backbones, COMFYCLAW achieves the best average image-generation evaluation score across all six agent configurations, outperforming a verifier-only baseline without skill evolution. Human annotations further show that annotators prefer COMFYCLAW over variants without skill evolution. Our results suggest that skill evolution is an effective mechanism for improving agent reliability and performance in recurring visual workflow construction.", "published": "2026-07-02", "updated": "2026-07-02", "authors": [ "Zongxia Li", "Dawei Liu", "Fuxiao Liu", "Yuhang Zhou", "Xiyang Wu", "Jingxi Chen", "Jing Xie", "Xiaomin Wu", "Lichao Sun" ], "categories": [ "cs.AI", "cs.LG" ], "url": "https://arxiv.org/abs/2607.01709", "pdf_url": "https://arxiv.org/pdf/2607.01709", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2607.01916", "title": "ContextSniper: AntTrail's Token-Efficient Code Memory for Repository-Level Program Repair", "abstract": "Large language model agents can repair real repository issues, but they often spend large context budgets on whole-file reads, broad searches, and long terminal outputs where useful evidence is mixed with irrelevant code and logs. This paper presents ContextSniper, AntTrail's code-repair module for precision evidence selection in repository-level program repair, part of AntTrail's broader agent-memory engine. AntTrail is available at https://gitcode.com/datagallery/AntTrail. ContextSniper indexes code and action memory as three abstract levels, retrieves candidates with a hybrid ranker, filters long tool output through an intention-aware context gate, and returns compact evidence packets while keeping full source recoverable on demand. In a matched 50-task-per-condition comparison on SWE-bench Lite (same tasks, baseline vs.\\ ContextSniper), ContextSniper reduces total token use by 51.5% and logged cost by 36.4% for OpenClaw, and by 38.9% and 27.3% for Claude Code, with submitted-resolution rates essentially unchanged in both host-agent settings. In a separate five-task comparison, ContextSniper beats existing memory- and RAG-style integrations on token efficiency. These results suggest ContextSniper can substantially cut token and cost overhead for repository-level repair agents without a measurable loss in repair quality. The evaluation harness for this study is available at https://gitcode.com/lukchiwang/ContextSniper.", "published": "2026-07-02", "updated": "2026-07-09", "authors": [ "Chiwang Luk", "Matin Mohammad Najafi", "Zhifeng Jia", "Wei Yang", "Xiuchang Li", "Jinwei Zhu", "Yang Ren", "Lei Chen", "Gao Cong" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2607.01916", "pdf_url": "https://arxiv.org/pdf/2607.01916", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2607.01935", "title": "A-TMA: Decoupling State-Aware Memory Failures in Long-Term Agent Memory", "abstract": "Long term memory lets LLM agents act as persistent assistants, but user facts change. A useful memory system must know what is true now, what used to be true, and what changed. We study \\emph{ghost memory}, a state coordination failure in which old, current, and transition facts coexist in the memory bank, remain mixed during retrieval, and mislead the answer model. We argue that memory systems should be understood and optimized from three levels: bank maintenance, retrieval, and answer time resolution. We propose ATMA, a state aware overlay for existing memory systems. ATMA keeps superseded and transition records in the bank, builds evidence packets for the query's requested state view, and exposes current, historical, and transition labels to QA. We further call for decoupled evaluation of bank, retrieval, and answer level failures, since final QA accuracy can hide where ghost memory occurs. To make this failure measurable, we build LTP (LoCoMo Temporal Plus), a conflict heavy benchmark for ghost memory, and evaluate on LoCoMo for long conversation generalization. On LTP, Graphiti+ATMA improves conflict accuracy by 0.240 absolute over Graphiti. On LoCoMo, Graphiti+ATMA raises temporal F1 from 0.0295 to 0.1705. The gains are host dependent, but they indicate that explicit state roles can reduce memory failures hidden by final QA accuracy.", "published": "2026-07-02", "updated": "2026-07-08", "authors": [ "Zitong Shi", "Yixuan Tang", "Anthony Kum Hoe Tung" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2607.01935", "pdf_url": "https://arxiv.org/pdf/2607.01935", "matched_queries": [ "agent-memory", "long-term-memory" ] }, { "arxiv_id": "2607.03726", "title": "SelfMem: Self-Optimizing Memory for AI Agents", "abstract": "While current AI agents support increasingly long context windows, tool use, and skill execution for long-horizon tasks, they still require memory systems to effectively leverage historical experience. Existing memory frameworks typically rely on fixed storage, retrieval, and summarization mechanisms, which can be rigid across different tasks and often require manual tuning. To address this limitation, we propose SelfMem, a self-optimizing memory framework. Inspired by prior work on self-improving AI, we follow the principle of \"teaching an agent to fish rather than giving it a fish.\" Instead of forcing the model to follow a predefined memory strategy or format, SelfMem provides an environment with memory tools and feedback signals that allow the agent to explore, evaluate, and refine its own memory strategy. Our results show that SelfMem consistently outperforms retrieval, compression, and agent-memory baselines on BEAM across conversation scales from 100K to 1M tokens. Compared with the strongest baseline, SelfMem improves the official score by 48.7%, 40.8%, and 41.9% at 100K, 500K, and 1M, respectively. Further question-type analysis shows broad robustness across diverse memory demands, and our optimization study shows that model-guided strategy refinement further improves performance.", "published": "2026-07-04", "updated": "2026-07-04", "authors": [ "Shu Yang", "Junchao Wu", "Derek F. Wong", "Di Wang" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2607.03726", "pdf_url": "https://arxiv.org/pdf/2607.03726", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2607.04089", "title": "PLACEMEM: Toward a Compute-Aware Memory Plane for Lifelong Agents", "abstract": "Lifelong agents need more than larger context windows and better retrieval. They need memories that can persist, evolve, and be corrected without forcing the serving stack to recompute the same history on every turn or silently reuse stale runtime state. We present PLACEMEM as a systems position on lifelong-agent memory, instantiated by an executable control-plane prototype. The central claim is that agent memory should be represented as versioned capsules that unify semantics, provenance, validity, and reusable runtime state under one correction-aware identity. In the current prototype, capsules drive prompt-level text retrieval, KV-aware routing, and cascading invalidation over live streamed backends; prospective layer-frontier replay is intentionally framed as a deeper integration agenda rather than a claimed engine feature. We describe a vLLM-first prototype with persistent capsule state, concurrency-safe invalidation, an OpenAI-compatible routing sidecar, a typed metadata contract, and a benchmark harness that measures live first-token latency, reuse, and post-correction behavior. The result is both an executable artifact that demonstrates correction-aware control-plane behavior today and a concrete roadmap for replay-aware serving integration in future lifelong-agent systems.", "published": "2026-07-05", "updated": "2026-07-05", "authors": [ "Sukanta Ganguly" ], "categories": [ "cs.AI" ], "url": "https://arxiv.org/abs/2607.04089", "pdf_url": "https://arxiv.org/pdf/2607.04089", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2607.04391", "title": "Memory-Orchestrated Semantic System (MOSS): An Auditable Agentic Memory Architecture", "abstract": "Long-term memory remains a structural weakness of AI agents. The dominant approach, retrieval-augmented generation (RAG), relies on embedding-based similarity search, which is opaque by construction, difficult to audit, and bounded by the theoretical limits of vector representations. We present the Memory-Orchestrated Semantic System (MOSS), an agentic memory architecture in which the agent drives retrieval over a structured relational database. MOSS is model-agnostic, storage-agnostic, and API-agnostic: it runs on any relational engine, connects to any LLM provider (or to deterministic non-LLM processes), and deploys on any infrastructure, local or cloud. Its retrieval execution is symbolic and reproducible (once a query is formulated, no LLM participates in the retrieval loop) and every step of the system, from indexing to answer formulation, is logged and inspectable, making MOSS auditable by construction. Rather than imposing an external ontology, MOSS derives its conceptual vocabulary from the corpus itself. We report on a longitudinal deployment unique in the agentic-memory literature: a year of continuous production over an individual scholar's working corpus--a conversational corpus reaching back to October 2024 (some 44 million tokens, retroactively indexed) comprising 110,183 segments, alongside 163,494 catalogued documents, 569 inductively derived concepts, 322,662 concept annotations, and eleven metadata graphs totaling approximately five million relations--across four successive infrastructure generations. While the present case is that of a single researcher, the architecture is in no way specific to one person: it serves a team, an institution, or any entity that accumulates knowledge over time. We argue that auditable, sovereign, structurally unbounded memory is a precondition for AI agents intended to accompany a person or an organization over years rather than sessions.", "published": "2026-07-05", "updated": "2026-07-05", "authors": [ "Serge Lacasse", "Jérémie Hatier", "Alex Baker" ], "categories": [ "cs.CL" ], "url": "https://arxiv.org/abs/2607.04391", "pdf_url": "https://arxiv.org/pdf/2607.04391", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2607.05029", "title": "Your Agent's Memories Are Not Its Own: Forged Reasoning Attacks on LLM Agent Memory and Defenses", "abstract": "Persistent memory has enabled large language model (LLM) agents to store factual knowledge, prior decisions, reasoning histories, tool usage information, and context. While this has improved the agent's functionality and continuity across tasks, it has also introduced a new attack surface: the agent's own reasoning history. In this paper, we introduce the Forged Amplifying Rationale Memory Attack (FARMA), which poisons an agent's remembered reasoning rather than its factual knowledge. It inserts forged reasoning traces using evasive language that bypasses keyword-based defenses, then amplifies them through self-referential reinforcement that defeats consensus-based defenses. To address FARMA, we introduce SENTINEL, a layered defense pipeline to detect forged reasoning entries. Its central component is the Reasoning Guard that structurally analyzes candidate entries for forgery using five weighted signals. We evaluate FARMA and SENTINEL across multiple agents and different LLM models with 50 trials and show that FARMA achieves an attack success rate of up to 100% under baseline conditions and is capable of defeating defense mechanisms like keyword filter and A-MemGuard. Our evaluation also shows that SENTINEL reduces FARMA's attack success rate to as low as 0% with no false positives observed across 326 benign agent traces. Our work demonstrates the need to protect not only an agent's retrieved content but also the integrity of its reasoning history.", "published": "2026-07-06", "updated": "2026-07-06", "authors": [ "Neeraj Karamchandani", "Piyush Nagasubramaniam", "Sencun Zhu", "Dinghao Wu" ], "categories": [ "cs.CR", "cs.AI" ], "url": "https://arxiv.org/abs/2607.05029", "pdf_url": "https://arxiv.org/pdf/2607.05029", "matched_queries": [ "agent-memory", "long-term-memory" ] }, { "arxiv_id": "2607.05577", "title": "Narrative World Model: Narratology-Grounded Writer Memory for Long-Form Fiction", "abstract": "Long-form fiction writers need memory that answers multi-hop questions about evolving story state: who knows a secret and when they learned it, whether an event preceded the narration that revealed it, whether a setup paid off, and how a relationship shifted. General-purpose retrieval and agent-memory systems represent entities and facts but not the narratological structure these questions turn on, so they surface the wrong evidence or none at all. We introduce the Narrative World Model (NWM), a writer-memory system that pairs a narratology-grounded typed temporal-state graph with query-conditioned hybrid retrieval. To measure memory rather than the answerer, we read every system through a single held-constant Opus 4.8 reader over only that system's chapter-safe evidence, on a reproducible public corpus and a validated multi-hop benchmark, and we compare against the strongest existing temporal-knowledge-graph agent-memory framework, Graphiti/Zep (Rasmussen et al., 2025). NWM substantially and significantly outperforms this baseline on multi-hop narratological QA across both corpora, and far exceeds GraphRAG and flat retrieval. The advantage is representational rather than an artifact of extraction: it survives rebuilding the baseline with NWM's own extractor, and traces to its narratology-grounded structure and query-conditioned retrieval, not to graph size or extractor quality.", "published": "2026-07-06", "updated": "2026-07-06", "authors": [ "Mohammad Saifullah", "Thomas Kornmaier", "Taaha Kazi", "Vasu Sharma", "Aditya Sanjiv Kanade", "Aanand Kumar Yadav" ], "categories": [ "cs.AI", "cs.CL", "cs.IR" ], "url": "https://arxiv.org/abs/2607.05577", "pdf_url": "https://arxiv.org/pdf/2607.05577", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2607.06595", "title": "When Agents Remember Too Much: Memory Poisoning Attacks on Large Language Model Agents", "abstract": "Personal AI agents powered by large language models can reason and act using available tools to access emails, manage calendars, and push code to remote repositories, all with minimal oversight. When augmented with long-term memory, an agent can recall specific details relevant to the current task, reducing the need for large context windows. Currently, long-term memory agents tend to fall into two distinct domains: conversational and action-planning agents. Personal assistant agents sit at the convergence of these two domains and handle sensitive information while interacting with untrusted information sources, creating previously unaccounted security vulnerabilities. In this work, we introduce the novel attack vector, GhostWriter, which exploits current memory subsystems in tool-using personal agents to poison their memory store. GhostWriter operates in two phases: injection, where an adversary sends a hidden attack payload to the target agent; and activation, in which the poisoned memory is retrieved. We show that GhostWriter achieves near-universal injection rates of approximately 98% and a high average activation rate of approximately 60% against state-of-the-art agents. This attack is possible due to the lack of security-focused memory governance. In response, we propose Agentic Memory Sentry (AM-Sentry), which leverages two mitigation techniques: a memory-saving policy and a memory-retrieval screen. Our experiments show that AM-Sentry dramatically reduces GhostWriter's success rate while preserving agent utility.", "published": "2026-07-06", "updated": "2026-07-06", "authors": [ "George Torres", "Sharad Shrestha", "Satyajayant Misra" ], "categories": [ "cs.CR", "cs.AI" ], "url": "https://arxiv.org/abs/2607.06595", "pdf_url": "https://arxiv.org/pdf/2607.06595", "matched_queries": [ "agent-memory", "long-term-memory" ] }, { "arxiv_id": "2607.06195", "title": "LogicHunter: Testing LLM Agent Frameworks with an Agentic Oracle", "abstract": "Large Language Model (LLM) agent frameworks such as LangChain, LlamaIndex, and CrewAI have become critical infrastructure powering production AI systems, yet they remain severely under-tested due to fundamental challenges in automated testing. Unlike traditional software, where crashes serve as reliable oracles, defects in these pure Python frameworks manifest as ordinary exceptions or silent semantic failures, creating profound oracle ambiguity. This problem is exacerbated by strict type governance through Pydantic schemas and complex protocol requirements that cause existing fuzzers to generate overwhelming invalid inputs, while traditional test generators produce only trivial cases with weak regression assertions. We present LogicHunter, a fuzzing framework that addresses both the generation and oracle challenges through active specification-aware testing. LogicHunter employs specification-driven generation that systematically fuses formal type constraints with authentic usage patterns from real-world repositories, synthesizing inputs that are valid by construction yet semantically extreme, equipped with behavioral probes to expose silent failures. To resolve oracle ambiguity, we introduce the Agentic Oracle, which transcends passive classification by actively retrieving documentation, navigating source code, and inspecting runtime states through a ReAct-based architecture with Dual-Layer State Management and Dual-Stream Memory. Evaluated on three widely deployed frameworks, LogicHunter discovered 40 previously unknown bugs with 30 confirmed and 26 fixed by developers, while state-of-the-art baselines reported no bugs as final findings. The Agentic Oracle achieves 91.17% precision, surpassing the best passive approach at 29.27% by 61 percentage points.", "published": "2026-07-07", "updated": "2026-07-07", "authors": [ "Minghui Long", "Yanjie Zhao", "Haoyu Wang" ], "categories": [ "cs.SE" ], "url": "https://arxiv.org/abs/2607.06195", "pdf_url": "https://arxiv.org/pdf/2607.06195", "matched_queries": [ "memory-evaluation" ] }, { "arxiv_id": "2607.07108", "title": "Seeing and Reflecting: Multimodal Memory-Enhanced Agent Collaboration for Recommendation", "abstract": "Large language model (LLM)-based agentic recommender systems show promise in modeling user preferences through natural-language reasoning, yet they remain limited by text-centric inputs and coarse-grained memory updates, making agents prone to missing visual evidence, semantic noise, and preference drift. To address these limitations, we propose MMEACR, a Multimodal Memory-Enhanced Agent Collaboration framework for recommendation. MMEACR introduces a dual-track memory architecture that separates interpretable agent reasoning from fine-grained multimodal matching. In the reasoning track, collaborative User and Item Memory Agents maintain persistent multimodal memories and update them through an attribute-guided reinforcement-and-reflection mechanism. In the matching track, a decoupled multi-modal embedding memory is built from raw interaction narratives and item images to preserve detailed cross-modal signals beyond structured memory updates. The two tracks are integrated through weighted Reciprocal Rank Fusion to produce robust and interpretable rankings. Experiments on three real-world domains show that MMEACR achieves strong overall performance against competitive LLM-based and agent-based baselines, with notable gains in visually grounded recommendation scenarios.", "published": "2026-07-08", "updated": "2026-07-08", "authors": [ "Hao Cong", "Huizu Lin", "Zihan Wang", "Chengkai Huang", "Quan Z. Sheng", "Lina Yao" ], "categories": [ "cs.IR" ], "url": "https://arxiv.org/abs/2607.07108", "pdf_url": "https://arxiv.org/pdf/2607.07108", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2607.08032", "title": "What to Keep, What to Forget: A Rate--Distortion View of Memory Compaction in LLMs and Agents", "abstract": "Large language models, and the agents built on them, spend an ever-growing share of their compute and memory on remembering: caching attention keys and values, carrying long prompts, maintaining recurrent state, and storing what happened in previous turns and sessions. Because none of this memory is free, four largely separate research communities have each learned to compact it. They evict or quantize the KV cache, prune or distill prompts, bound architectural state, and consolidate agent memory. We argue that these are instances of one problem: a rate--distortion decision about what context-derived information to retain versus discard, at what fidelity, under a resource budget, so as to preserve downstream task utility. We make this lens precise with a single compaction objective and a layer-agnostic lower bound, use it to build a seven-axis taxonomy that classifies methods from across the stack uniformly, and use it to transfer mechanisms between layers that have never been connected, from serving-stack KV management to agent long-term memory. Two patterns hold across the survey. At every layer the signal that decides what to keep is attention magnitude or recency, and it fails in the same way everywhere, by discarding, before the query is known and with no way to undo it, information the query later needs. And while compression is measured carefully on single-turn long context, the repeated compaction that agents actually perform is almost never measured, and no benchmark holds one budget axis across all the layers at once. We turn both observations into a benchmark proposal, a small reference experiment, and a set of compaction-aware design principles, and we map the open problems.", "published": "2026-07-09", "updated": "2026-07-09", "authors": [ "Ashwin Gerard Colaco", "Nada Lahjouji" ], "categories": [ "cs.LG" ], "url": "https://arxiv.org/abs/2607.08032", "pdf_url": "https://arxiv.org/pdf/2607.08032", "matched_queries": [ "agent-memory" ] }, { "arxiv_id": "2607.08716", "title": "Remember When It Matters: Proactive Memory Agent for Long-Horizon Agents", "abstract": "In long-horizon tasks, decision-relevant state is often scattered across an expanding trajectory, while the action agent must surface it and act. As trajectories grow, task requirements, environment facts, prior attempts, diagnoses, and open subgoals can be buried in the context window or pushed beyond it, failing to influence decisions when needed. We call this failure mode \"behavioral state decay\". We study memory as an active intervention mechanism rather than passive retrieval. A separate memory agent runs alongside an unmodified action agent, updating a structured memory bank from the recent trajectory and deciding whether to inject a memory-grounded reminder or remain silent. The module is plug-and-play with frontier action agents and existing agent harnesses. Across Terminal-Bench 2.0 and $τ^2$-Bench, it improves pass@1 for both weaker and stronger action agents, with gains of +8.3 pp on Terminal-Bench and +6.8 pp on $τ^2$-Bench. Ablations show that selective intervention outperforms passive bank exposure, always-on injection, advisor-only guidance, and general retrieval. As an early step toward open-weight memory policies, we train Qwen3.5-27B on SETA using SFT and GRPO, improving validation reward and achieving partial transfer to Terminal-Bench.", "published": "2026-07-09", "updated": "2026-07-09", "authors": [ "Yifan Wu", "Lizhu Zhang", "Yuhang Zhou", "Mingyi Wang", "Bo Peng", "Serena Li", "Xiangjun Fan", "Zhuokai Zhao" ], "categories": [ "cs.AI", "cs.CL" ], "url": "https://arxiv.org/abs/2607.08716", "pdf_url": "https://arxiv.org/pdf/2607.08716", "matched_queries": [ "agent-memory" ] } ] }