让大模型动态拆解网页任务并实时重规划，显著提升复杂操作成功率。
WebDART: Dynamic Decomposition and Re-planning for Complex Web Tasks
- 将复杂任务分解为导航、信息提取、执行三类子任务，逐个专注解决
- 在新页面出现时持续重规划，减少14.7步导航，成功率达93.6%
- 适合需要长程推理与多步操作的自动化网页任务场景
大型语言模型代理在简单网页任务（如打开商品页或提交表单）上已表现良好，但在需要长程导航、大规模信息抽取及受限条件下的推理任务中仍显不足。本文提出WebDART框架，使单一大模型可高效处理此类复杂任务。该框架（i）动态将目标分解为导航、信息提取与执行三类子任务，使模型分阶段聚焦单一技能；（ii）在浏览过程中持续重规划分解策略，利用新发现的筛选条件或捷径路径，避免重复探索。在WebChoreArena上的评估显示，WebDART相较此前最优模型，成功率最高提升13.7个百分点，且在更简单的WebArena基准上保持相当性能，同时任务平均减少14.7次导航步骤。
原文摘要 · Abstract (English)
Large language model (LLM) agents are becoming competent at straightforward web tasks, such as opening an item page or submitting a form, but still struggle with objectives that require long horizon navigation, large scale information extraction, and reasoning under constraints. We present WebDART, a general framework that enables a single LLM to handle such complex chores. WebDART (i) dynamically decomposes each objective into three focused subtasks: navigation, information extraction, and execution, so the model concentrates on one skill at a time, and (ii) continuously replans the decomposition as new webpages are revealed, taking advantage of newly discovered filters or shortcuts and avoiding redundant exploration. Evaluated on WebChoreArena, WebDART lifts success rates by up to 13.7 percentage points over previous SOTA agents, while matching their performance on the easier WebArena suite and completing tasks with up to 14.7 fewer navigation steps.
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。