arXiv cs.AI· Genliang Zhu, Xiaobo Chen, Ziyuan Wang, Chu Wang·· 1 天前AI 评分31
面向工具结果模糊与部分执行的 AI 智能体证据引导恢复方法
Evidence-Guided Recovery for AI Agents with Ambiguous Tool Outcomes and Partial Execution
AI 导读
研究提出证据引导的任务恢复方法，应对工具超时导致的模糊结果与部分执行，通过共享可接受性检查处理密钥过期、授权时序与依赖关系，并用终态证书区分真实目标达成与公开支持的安全完成。在 19 个任务家族的 58 个配对单元中，规划建议认证 30 次完成，证据提示仅 9 次，相差 36.2 个百分点（精确家族符号翻转 p=0.0078），提案数 272 对 399，全部 116 条轨迹通过独立重放。
正文 · 原文
Abstract:Tool timeouts can leave AI agents facing ambiguous tool outcomes and partial execution: an action may have failed, committed or remained in flight while other workflow steps have already completed. We formulate evidence-guided task recovery over approved obligations, provider capabilities and states compatible with public evidence. A shared admissibility check handles key expiry, authorization timing, dependencies and residual effects; a terminal certificate separates actual goal satisfaction from publicly supported safe completion. The known-model implementation combines standard observation-contingent planning with a separately implemented effect-history adjudicator. In a finite study of 37 assignments, planning and three active controls each certify 28 completions; four saved synthetic fee units incur additional computation. A preregistered assignment-held-out comparison then evaluates planning advice against evidence prompting with two local model bundles. Across 58 paired units in 19 previously developed task families, advice certifies 30 completions versus 9, a 36.2-percentage-point difference (exact family sign-flip p=0.0078), using 272 proposals versus 399. All 116 trajectories pass independent replay. Both models exhibit a positive completion difference. The comparison measures the evaluated advice interface under a common shield; search-versus-rule advice requires a separate matched comparison. Unknown and handoff outcomes remain explicit, with guarantees tied to the declared provider capabilities rather than arbitrary third-party exactly-once execution.
| Comments: | 24 pages, 1 figure, and 10 tables. Includes a finite recovery model, model-facing confirmation study with two local-model bundles, independent replay, and an explicit provider-capability scope |
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.41453 [cs.AI] |
| (or arXiv:2610.41453v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.41453 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Chu Wang [view email]
[v1]
Fri, 30 Oct 2026 08:55:55 UTC (36 KB)
来源：arXiv cs.AI · arxiv.org