Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2601.21714 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866917492795899904 |
|---|---|
| author | Wang, Kaixiang Lin, Yidan Lou, Jiong Zhou, Zhaojiacheng Suvonov, Bunyod Li, Jie |
| author_facet | Wang, Kaixiang Lin, Yidan Lou, Jiong Zhou, Zhaojiacheng Suvonov, Bunyod Li, Jie |
| contents | 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\%. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_21714 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | E-mem: Multi-agent based Episodic Context Reconstruction for LLM Agent Memory Wang, Kaixiang Lin, Yidan Lou, Jiong Zhou, Zhaojiacheng Suvonov, Bunyod Li, Jie Artificial Intelligence 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\%. |
| title | E-mem: Multi-agent based Episodic Context Reconstruction for LLM Agent Memory |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2601.21714 |