发表机构
The Hong Kong Polytechnic University; Platform and Content Group, Tencent; Soochow University（香港理工大学; 腾讯平台与内容事业群; 苏州大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本文提出SkillMentor，将盲点诊断作为独立于执行的可学习智能体能力，通过强化学习训练导师策略，在AppWorld和BFCLv3上使执行器性能平均提升44.2%，实现无需更新执行器的智能体自我进化。
AI 中文摘要
智能体自我进化主要聚焦于学习如何行动，却忽略了同等重要的能力：发现智能体未知内容的能力。现有方法通常假设故障发现已完成，专注于识别故障后的修复。本文探究盲点诊断本身是否可学习，将诊断视为与执行分离的智能体能力，排除执行器适配和人工监督两种替代进步来源。在此约束下，性能无法通过执行器更新或标注示例提升，所有改进需源于学习到的诊断能力。本文提出SkillMentor，通过强化学习训练导师策略，生成诊断任务、识别反复出现的故障模式并将其整理为可复用的修正技能。在AppWorld和BFCLv3数据集上，SkillMentor使执行器性能平均提升44.2%。这些结果表明，盲点诊断是可学习的能力，可实现无需更新执行器权重或依赖人工整理数据的自我进化。
英文摘要
Agent self-evolution has primarily focused on learning how to act, while overlooking an equally important capability: learning to discover what an agent does not know. Existing approaches typically assume that failure discovery is given, focusing on how to repair failures once they are identified. We ask whether blind-spot diagnosis itself can be learned. We thus study diagnosis as an agent capability separate from execution, and exclude two alternative sources of progress: executor adaptation and human supervision. Under these constraints, performance cannot improve through executor updates or annotated examples, forcing all improvements to originate from the learned diagnostic capability. We propose SkillMentor, which trains a Mentor policy via reinforcement learning to generate diagnostic tasks, identify recurrent failure modes, and curate them into reusable corrective skills. Across AppWorld and BFCLv3, SkillMentor improves executor performance by an average of 44.2%. These results suggest that blind-spot diagnosis is a learnable capability, enabling self-evolution without updating executor weights or relying on human-curated data.
