PatchScout：基于信息觅食的主题式网络数据采集
原标题：PatchScout: Thematic Web Data Collection via Information Foraging
AI 摘要
研究者提出 PatchScout，一个多智能体框架，用于解决「主题式网络数据采集」问题：即从分散、结构上互不连通的网页区域中收集共享同一主题的文档集合。该方法将任务建模为长时程信息觅食问题，通过一组 patch 切换策略在固定预算下平衡局部开发与新区域探索。在四个实时网络主题上的实验显示，PatchScout 在产出量和域名覆盖率上均优于基于遍历和搜索增强推理智能体的基线；在社会科学应用场景中，它还发现了 102 个此前未被识别的实体。
正文节选
PatchScout: Thematic Web Data Collection via Information Foraging Abstract Researchers often need web corpora that are not answers to a single query, but reusable collections of documents spanning many decentralized sources. We refer to this problem as thematic web data collection: assembling semantically relevant documents that share a common theme but are scattered across heterogeneous and structurally disconnected regions of the web. Existing approaches, including traversal-based methods and