arXiv cs.AI· Shixuan Li, Wei Yang, Peiyu Zhang, Anzhe Cheng, Heng Ping, Xiaole Zhang, Jesse Thomason, Xuezhe Ma, Yue Zhao, Paul Bogdan·· 12 小时前AI 评分38
QEMem 框架：从相关记忆到充分证据，重新思考 LLM 智能体的记忆利用
From Relevant Memories to Sufficient Evidence: Rethinking Memory Utilization in LLM Agents
AI 导读
研究者提出 QEMem 框架，将 LLM 智能体的记忆利用重构为构建查询专属证据集，而非依赖单条记忆的相关性。
正文 · 原文
Abstract:Long-term memory enables LLM agents to reuse past experiences, yet retrieving relevant memories does not necessarily provide sufficient evidence for answering a query. Essential information may remain fragmented across memory records, obscured by abstraction, or disconnected from the relations required for reasoning. We identify this relevance-sufficiency gap and propose QEMem, a framework that reformulates memory utilization as the construction of query-specific evidence sets, shifting the focus from individual retrieval relevance to collective evidential support. QEMem integrates three complementary mechanisms: query-guided discovery retrieves evidence across multiple memory granularities; source-grounded reification organizes it into structured ViewCards with explicit evidential roles and source references; and sufficiency-guided refinement selectively recovers missing information. Operating on existing memory stores without modifying memory formation, QEMem achieves 91.43% LLM-judge accuracy on LoCoMo with Claude Opus 4.5, outperforming the strongest evaluated baseline, MemR$^3$, by 7.6 percentage points and NEMORI by 10.7 points under a shared memory store. Further analyses demonstrate improved evidence coverage and strong performance on questions requiring cross-session evidence integration.
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.22668 [cs.AI] |
| (or arXiv:2610.22668v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.22668 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Shixuan Li [view email]
[v1]
Thu, 15 Oct 2026 21:34:17 UTC (321 KB)
来源：arXiv cs.AI · arxiv.org