跳到正文
arXiv cs.AI· \"Omer Faruk Akg\"ul, Shuli Jiang, Yuting Zhang, Yi Zhang, Xiaohan Fei, Wei Xia, Stefano Soatto·· 56 分钟前AI 评分51

Agent 技能过时问题研究:选择性维护让任务准确率最多提升 67pp

When Agent Skills Go Stale: Runtime Behavior and Selective Maintenance in Evolving Environments

AI 导读

arXiv 论文(arXiv:2611.22525)研究 Agent 技能状态在环境变化后过时的问题:四类任务和两个模型上,此前有益的技能可使任务表现下降超过 10 个百分点或执行成本增加近 40%。论文提出选择性维护方法,通过重放成功执行的检查来验证、修复并验证技能状态,无需额外模型调用,相比复用过期状态可将任务准确率提升最多 67pp、执行步骤减少最多 37%。

正文 · 原文

View PDF HTML (experimental)

Abstract:Agent skills use past experience and current observations to guide decisions and actions on related tasks. This experience is retained in skill state, whose reuse can amortize inference-time reasoning and environment interaction. When environment changes, parts of this state can become outdated while others remain useful. Across four task settings and two models, we compare skill reuse before and after environment changes. Previously beneficial skills can reduce task performance by over 10 percentage points or increase execution cost by nearly 40% relative to no skill. Task success can also remain unchanged while execution savings decline. These effects depend on which environment-specific details become outdated and the agent's ability to compensate using current observations. We introduce selective maintenance, which replays checks from successful executions without additional model calls. It verifies whether observed differences conflict with skill state, repairs affected content, and validates updates, rebuilding when local repair is insufficient. No description of the change is supplied. Relative to reusing stale state, maintenance improves task accuracy by up to 67pp and reduces execution steps by up to 37%. These results suggest that preserving a skill's inference time value as environments evolve requires selectively maintaining its state.
Subjects: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
Cite as: arXiv:2611.22525 [cs.AI]
  (or arXiv:2611.22525v1 [cs.AI] for this version)
  https://doi.org/10.48550/arXiv.2611.22525

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Ömer Faruk Akgül [view email]
[v1] Tue, 17 Nov 2026 10:18:46 UTC (656 KB)

来源:arXiv cs.AI · arxiv.org