arXiv：cs.AI· Zihao Lu, Zhihang Yuan, Litu Ou, Lei Shi·· 5 小时前
EGMEMORY：面向长程多参与者对话的证据链记忆框架
Propose, Verify, Commit: Evidence-Grounded Memory for Long-Horizon Multi-Actor Conversations
AI 导读
EGMEMORY 是一个免训练的智能体记忆框架，以证据链保留消息级证据与来源，并追踪状态的修订历史。它在 GroupMemBench 上达到 68.2%、EverMemBench 上达到 77.9%，分别超出最强基线 22.7 和 21.4 个百分点，在 LoCoMo 上取得 73.6%。该系统仅通过提示词与工具调用运行，无需针对记忆的策略训练。
正文
Abstract:Memory enables language model agents to use information beyond their context windows, but maintaining and accessing this information reliably remains difficult. Long-horizon conversational memory is especially challenging in multi-actor settings, where relevant evidence is distributed across participants and contexts, and previously established information may later be revised. We introduce EGMEMORY, a training-free agentic memory framework centered on evidence chains that preserve message-level evidence and provenance while tracking active states and their revision histories. At write time, adaptive memory inspection and an evidence-grounded propose-verify-commit protocol govern state updates. At read time, adaptive evidence navigation iteratively identifies the state and supporting evidence required for a query, using conversational structure to narrow the search space and lexical-semantic relevance to rank candidates. The system operates through prompting and tool use without memory-specific policy training. EGMEMORY achieves 68.2% on GroupMemBench and 77.9% on EverMemBench, outperforming the strongest evaluated baselines by 22.7 and 21.4 percentage points, respectively. It further reaches 73.6% on the dyadic LoCoMo benchmark, supporting its applicability beyond multi-actor conversations.
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2609.23465 [cs.CL] |
| (or arXiv:2609.23465v2 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2609.23465 arXiv-issued DOI via DataCite |
Submission history
From: Zhihang Yuan [view email]
[v1]
Sun, 20 Sep 2026 08:51:07 UTC (141 KB)
[v2]
Tue, 13 Oct 2026 11:17:52 UTC (178 KB)
来源：arXiv：cs.AI · arxiv.org