让智能体通过编程方式自动生成、验证和复用技能，提升网页任务执行效率。
Inducing Programmatic Skills for Agentic Tasks
- 用程序作为技能表示，动态生成并验证网页操作技能。
- 在WebArena上成功率提升23.5%，步骤减少10.7%-15.3%。
- 技能可跨网站复用，适应不同网站变化，适合复杂网页代理场景。
为完成如网页导航等常见数字任务，智能体需掌握搜索商品或规划行程等专项技能。本文提出代理技能诱导（ASI），使智能体能通过与网络环境交互，在线动态生成、验证并使用基于程序的技能。在WebArena基准测试中，ASI相较于静态基线和文本技能模型，成功率分别提升23.5%和11.3%，主要得益于程序化验证机制。同时，通过将基础操作（如点击）组合为高层技能（如搜索商品），步骤数减少10.7%-15.3%。在扩展的网页活动下，ASI仍保持高效与准确。进一步实验表明，诱导出的技能可在不同网站间有效复用，对不兼容技能也能及时更新以适应网站变化。
原文摘要 · Abstract (English)
To succeed in common digital tasks such as web navigation, agents must carry out a variety of specialized tasks such as searching for products or planning a travel route. To tackle these tasks, agents can bootstrap themselves by learning task-specific skills online through interaction with the web environment. In this work, we demonstrate that programs are an effective representation for skills. We propose agent skill induction (ASI), which allows agents to adapt themselves by inducing, verifying, and utilizing program-based skills on the fly. We start with an evaluation on the WebArena agent benchmark and show that ASI outperforms the static baseline agent and its text-skill counterpart by 23.5% and 11.3% in success rate, mainly thanks to the programmatic verification guarantee during the induction phase. ASI also improves efficiency by reducing 10.7-15.3% of the steps over baselines, by composing primitive actions (e.g., click) into higher-level skills (e.g., search product). We then highlight the efficacy of ASI in remaining efficient and accurate under scaled-up web activities. Finally, we examine the generalizability of induced skills when transferring between websites, and find that ASI can effectively reuse common skills, while also updating incompatible skills to versatile website changes.
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。