让AI agent像搭积木一样自动升级技能，提升效率38%、减少27.7%步骤。
SkillPyramid: A Hierarchical Skill Consolidation Framework for Self-Evolving Agents
- 构建分层技能结构，让旧技能可复用、新技能可自动生成
- 在多个任务平台测试中，平均奖励提升38.0%，执行步骤减少27.7%
- 适合长期进化型AI系统研发者，尤其关注技能复用与泛化
近期的AI代理能灵活调用技能解决复杂任务，但其长期改进受限于缺乏系统性的技能构建、积累与迁移机制。由于缺乏统一的技能整合框架，代理在不同任务中重复构建相似能力，难以将经验转化为可复用资产，且难以将特定任务技能泛化到新场景。为此，我们提出SkillPyramid，一种基于分层技能拓扑的技能整合框架，通过自演化机制实现任务执行过程中的技能组合、验证与融合。在ALFWorld、WebShop和ScienceWorld三个平台，使用四种骨干模型进行实验，结果表明，SkillPyramid使平均奖励提升38.0%，执行步骤减少27.7%。整体上，该方法将技能集合从静态资源池转变为动态演化系统。
原文摘要 · Abstract (English)
Recent AI agents can flexibly invoke skills to solve complex tasks, but their long-term improvement is fundamentally constrained by a lack of systematic skill construction, accumulation, and transfer. In particular, without a unified framework for skill consolidation, agents tend to redundantly construct similar capabilities across different tasks, are unable to effectively transform experience into reusable assets, and struggle to generalize task-specific skills to novel scenarios. To address this limitation, we propose SkillPyramid, a skill consolidation framework that reuses existing skill experience for broader task generalization. Operating on a hierarchical skill topology, SkillPyramid further introduces a self-evolution mechanism that enables agents to compose, validate, and incorporate new skills during task execution. Experiments on ALFWorld, WebShop, and ScienceWorld across four backbone models show that SkillPyramid substantially increases the average reward by 38.0% and reduces execution steps by 27.7%. Overall, our method transforms a skill collection from a static resource pool into a dynamic evolution system.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。