arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2601.21714cs.AI
收藏

E-mem:基于多智能体的事件上下文重建用于LLM智能体记忆

E-mem: Multi-agent based Episodic Context Reconstruction for LLM Agent Memory

National University of Singapore(新加坡国立大学);Tsinghua University(清华大学)

Kaixiang Wang, Yidan Lin, Jiong Lou, Zhaojiacheng Zhou, Bunyod Suvonov, Jie Li

AI总结:

E-mem通过多智能体协同重建事件上下文,提升LLM在长时序任务中的推理能力,实现54%的F1分数,优于现有方法7.75%,并降低70%的token成本。

摘要:

大型语言模型(LLM)智能体向系统2推理发展,需要在长时空中保持严格的逻辑完整性。然而,主流的记忆预处理方法因破坏性去上下文化而不足。本文提出E-mem框架,从记忆预处理转向事件上下文重建。受生物酶体启发,E-mem采用异构分层架构,多个助手智能体维护未压缩的记忆上下文,而中央主智能体负责全局规划。与被动检索不同,我们的机制使助手能在激活段内本地推理,提取上下文感知的证据后再聚合。在LoCoMo基准测试中,E-mem实现了超过54%的F1分数,优于现有最佳方法GAM 7.75%,同时降低token成本超过70%。

英文摘要:

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\%.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑