agent/papers/items/2026-2606-06090-beyond-semantic-organization-memory-as-execution-state-management-for-long-horiz.md

Paper: Beyond Semantic Organization: Memory as Execution State Management for Long-Horizon Agents

---

type: paper title: "Beyond Semantic Organization: Memory as Execution State Management for Long-Horizon Agents" authors: Yaoqi Chen, Haibin Lai, Yuru Feng, Chuyu Han, Qianxi Zhang, Baotong Lu, Menghao Li, Xinjiang Wang, et al. year: 2026 venue: arXiv url: https://arxiv.org/abs/2606.06090 code_url: source: arxiv collected_at: 2026-07-08 published_at: 2026-06-04 updated_at: 2026-06-04 status: skimmed relevance: high topics:

• computer-use
• memory
• planning
• rag
• tool-use methods:
• hierarchical-execution-state-tree
• grow-compress-maintain-revise
• error-branch-isolation benchmarks:
• MemoryArena models:

datasets:

• cs.AI related_concepts:
• execution-state
• long-horizon-agent related_jobs:

related_experiments:

• KC-001-agent-memory-pilot related_projects:
• learning/agent-memory collection_score: 16 collection_queries: agent-memory, rag-agent

---

One-line Takeaway

长程任务需要维护 root-to-current 执行状态并隔离错误分支，而不是从相似记录中临时拼装上下文。

Pilot Skim

• problem: 相似度组织会割裂执行依赖，并把有效和错误轨迹重新混入同一上下文。
• method: 两层 state tree 保存 action-observation 与 subgoal summary，通过 Grow、Compress、Maintain、Revise 管理当前路径。
• evidence: MemoryArena 上相对 baselines 平均提高 7.8 至 20.4pp，并比长上下文减少 55.1% token。
• boundary: 证据针对 interdependent long-horizon tasks，不直接代表事实问答或个性化 memory。

Used In

• Agent Memory evidence matrix

Links

• arXiv: https://arxiv.org/abs/2606.06090