arXivDaily arXiv每日学术速递 周一至周五更新
​通知:联邦节假日调休:哥伦布日 / 原住民日(Columbus Day / Indigenous Peoples' Day),今日(13日)无更新。
arXiv 2607.01874cs.AIcs.CL

SkillCoach: 用于评估和增强智能体技能使用的自演化评分准则

SkillCoach: Self-Evolving Rubrics for Evaluating and Enhancing Agentic Skill-Use

HKUST(GZ)(香港科技大学(广州))

Jiayin Zhu, Kelong Mao, Yudong Guo, Dengbo He, Sulong Xu, Simiu Gu, Yutao Yue

AI总结:

提出SkillCoach框架,通过自演化过程评分准则从技能选择、遵循、组合和反思四维度评估智能体轨迹,并利用演化准则筛选高质量训练数据,提升技能使用能力。

摘要:

技能正在成为LLM智能体的可重用操作层,编码SOP、领域规则、工具工作流、脚本和验证例程。在现实技能库中,重叠的技能使得可靠的技能使用变得困难。最终验证器的成功对于评估和训练都过于粗糙,因为智能体可能通过试错选择干扰技能、跳过必要步骤、错误组合工作流或省略最终检查。我们引入SkillCoach,一个自演化评分准则框架,用于评估和增强智能体技能使用。SkillCoach从真实轨迹中推导出基于技能的过程评分准则,并沿四个维度评估轨迹:技能选择、技能遵循、技能组合和基于技能的反思。它将外部验证器作为独立的结果信号,使过程质量与偶然的任务成功区分开来。演化的评分准则进一步作为过程监督,用于选择高质量的训练轨迹。实验表明,演化评分准则显著提高了评估质量,揭示了被最终准确性隐藏的失败,并为增强智能体技能使用提供了比仅结果过滤更强的监督信号。

英文摘要:

Skills are becoming a reusable operational layer for LLM agents, encoding SOPs, domain rules, tool workflows, scripts, and validation routines. In realistic skill repositories, overlapping skills make reliable skill-use difficult. Final verifier success is too coarse for both evaluation and training, since an agent may pass through trial and error while selecting distractor skills, skipping required steps, composing workflows incorrectly or omitting final checks. We introduce SkillCoach, a self-evolving rubric framework for evaluating and enhancing agentic skill-use. SkillCoach derives skill-grounded process rubrics from real rollouts and evaluates trajectories along four dimensions: skill selection, skill following, skill composition, and skill-grounded reflection. It keeps the external verifier as a separate outcome signal, allowing process quality to be distinguished from accidental task success. The evolved rubrics further serve as process supervision for selecting high-quality training trajectories. Experiments show that evolved rubrics substantially improve evaluation quality, expose failures hidden by final accuracy, and provide stronger supervision signals than outcome-only filtering for enhancing agentic skill-use.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑