EvoSkill-S1:用遗传算子让智能体技能库直接进化
EvoSkill-S1: Evolving Reusable Skills with Genetic-Inspired Operators
如果你在搭会积累经验的智能体,这篇挺值得看:它不是靠反复 rollout 慢慢攒技能,而是直接用 RL 训练一个技能进化器,配五个遗传算子改写旧技能,数学和代码任务上最高比外部基线好 20.39%。
论文提出 EvoSkill-S1,一个把可复用技能本身作为进化对象的 RL 框架。给定 hard-negative 失败样本后,它通过 revision、cross-revision、mutation、dual-reflection、crossover 五种遗传算子,从父技能生成候选子技能,并用下游任务奖励训练技能进化器,由 evaluator 决定哪些更新写回技能库。
机制上还包括 Skill Bank Fusion,把各算子单独形成的技能库合并以利用不同恢复模式,同时用 Procedural Skill Interface 支持向未见问题迁移。在数学和代码基准上,EvoSkill-S1 在 Qwen 和 GPT 系求解器下均优于初始技能库,相对提升最高 12.39%;与不含直接技能进化的最强外部基线相比,相对提升最高达 20.39%。
对做智能体的开发者来说,这条路线的意义在于技能更新从 rollout 间接积累和启发式反思,转为有奖励信号直接监督的受控进化,技能库的复用与跨任务迁移有了更明确的训练目标。
EvoSkill-S1: Evolving Reusable Skills with Genetic-Inspired Operators
Reusable skills allow language-model agents to accumulate and reuse problem-solving experience. However, most existing methods improve a skill bank only indirectly through rollout-driven accumulation and heuristic reflection, often resulting in shallow and weakly controlled updates. We present EvoSkill-S1, an RL-driven framework that instead treats reusable skills themselves as the direct objects of evolution. Given hard-negative failures, EvoSkill-S1 selects parent skills and transforms them into candidate child skills through five genetic-inspired operators: revision, cross-revision, mutation, dual-reflection, and crossover. Downstream task rewards train the skill evolver to generate more effective child skills, while evaluator-guided write-back selectively retains beneficial updates. Complementary Skill Bank Fusion consolidates operator-specific banks to exploit their distinct recovery patterns, while a Procedural Skill Interface supports transfer to unseen problems. Across all reported math and code benchmarks, EvoSkill-S1 improves the corresponding initialized skill banks under both Qwen and GPT-family solvers, with relative gains of up to 12.39%. Compared with the strongest external baseline without direct skill evolution, the largest relative improvement reaches 20.39%.
This report is categorized under AITOP's "Papers" desk. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include EvoSkill-S1, agents, skill reuse, RL.
Source attribution: curated from arXiv cs.AI (@Yujun Wang, Jiqiang Li, Yinuo Lu, Jinhe Bi, Aniri, Wenxuan Ye, Boliang Liu, Sikuan Yan, Shuning Wang, Xuebing Zhou, S{ö}ren Pirk, Hinrich Sch{ü}tze, Yunpu Ma).