arXiv cs.MA· Yiqi Wang, Jinqian Ju, Jiaqi Zhang, Zequn Sun, Yiqun Duan, Mingkai Zheng, Taotao Cai·· 12 小时前AI 评分31
执行溯源何时能提升 Agent 记忆检索？arXiv 论文提出预算化证据补全方法
When Does Execution Provenance Help Agent Memory Retrieval?
AI 导读
研究者将 Agent 记忆检索重构为预算化证据补全任务，用工具参数与输出构建源对齐的溯源单元，并在冻结的稠密检索分数上叠加零初始化残差 R-GCN 做图传播。
正文
Abstract:A language agent's execution history can exceed its context window, requiring its memory system to retrieve complete supporting evidence under a hard token budget. Evidence may span multiple execution events, yet conventional retrievers use fixed token windows and fixed-k metrics that reward individual fragments without showing whether the complete evidence set fits in context. Smaller windows reduce irrelevant text but scatter evidence across candidates, while flat-versus-graph comparisons can conflate candidate design with graph propagation. To address these limitations, we formulate agent-memory retrieval as budgeted evidence completion and score exact gold spans in shared source coordinates. We first construct source-aligned provenance units from tool arguments and outputs. We then apply a zero-initialized residual R-GCN to refine frozen dense-retrieval scores over typed provenance edges. We evaluate 2,000 span-grounded memory queries over 1,207 held-out execution-grounded ISETrace trajectories. With matched Dense-FT scoring, provenance units improve Full Support@2048 by 19.07 points over flat 512-token windows and remain 11.96 points above a per-metric oracle over four flat chunk sizes; the pattern also holds with cross-encoder scoring. Holding the candidates and seed scores fixed, graph propagation adds 4.55 points in Full Support@2048 (95% CI [2.98, 6.18]). This gain is concentrated when gold evidence spans multiple events; entity co-occurrence expansion produces no comparable benefit, and relation and topology controls confirm dependence on typed transformations and observed graph structure. Overall, source-aligned candidates address the dominant granularity trade-off, while graph-conditioned propagation adds a smaller, targeted benefit for distributed evidence.
| Comments: | 14 Pages, 4 Figures, 8 Tables |
| Subjects: | Multiagent Systems (cs.MA) |
| Cite as: | arXiv:2609.25913 [cs.MA] |
| (or arXiv:2609.25913v3 [cs.MA] for this version) | |
| https://doi.org/10.48550/arXiv.2609.25913 arXiv-issued DOI via DataCite |
Submission history
From: Jiaqi Zhang [view email]
[v1]
Tue, 22 Sep 2026 09:20:51 UTC (375 KB)
[v2]
Tue, 6 Oct 2026 01:58:15 UTC (374 KB)
[v3]
Sun, 25 Oct 2026 11:57:58 UTC (374 KB)
来源：arXiv cs.MA · arxiv.org