arXiv cs.AI· Runxuan Tang, Haoyu Gao, Yuyan Ding, Junyi Yao, Zihao Zheng, Zecheng Sheng, Liwei Hou·· 2 小时前AI 评分52
FGSE 框架提出证伪门控的自我进化方法解决 LLM 智能体记忆更新时机问题
When Should Agents Remember? Falsification-Gated Self-Evolution for LLM Agents
AI 导读
论文提出证伪门控自我进化框架 FGSE，回答 LLM 智能体经验教训何时应写入持久状态的问题：将候选教训表示为含前提条件、行为变化、预期效果和验证器的结构化假设，先在临时状态经目标迁移、证伪和归档回归探针测试，再按收益与风险决定提交、修正或拒绝。
正文 · 原文
Abstract:LLM agents increasingly operate in long-horizon, verifier-rich environments, but current self-improvement mechanisms often write episode-derived lessons directly into memory, making local success an unreliable signal for reusable and safe knowledge. This paper addresses the core question of when an experience-derived behavioral update should be admitted into persistent agent state. We propose Falsification-Gated Self-Evolution (FGSE), a verifier-grounded framework that represents each candidate lesson as a structured hypothesis with an explicit precondition, behavior change, expected effect, and verifier. FGSE applies the hypothesis only in a temporary state, tests it on target-transfer, falsification, and archived-regression probes, and then commits, refines, or rejects it according to measured gain and risk. Across web, app, tool-use, and long-term memory benchmarks, FGSE achieves strong task performance, including 39.8 WebArena success rate, 78.8 AppWorld average completion, and 0.724/0.487 Pass$^1$ on $\tau$-Retail/$\tau$-Airline, while reducing falsification failures to 5.0%, regression damage to 1.3%, and harmful committed updates to 3.6%. These results suggest that testing memory updates as hypotheses supports more reliable long-term adaptation.
| Subjects: | Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2611.14326 [cs.AI] |
| (or arXiv:2611.14326v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2611.14326 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Liwei Hou [view email]
[v1]
Tue, 10 Nov 2026 16:52:25 UTC (2,530 KB)
来源：arXiv cs.AI · arxiv.org