AttriSwarm：用代理集群实现跨页视觉归因
原标题：AttriSwarm: Orchestrating Agent Swarms for Cross-Page Visual Attribution
AI 摘要
该论文提出 AttriSwarm，一个 agent-swarm 框架，用于跨页视觉归因：由 leader 负责全局跨页推理，子代理在隔离上下文中并行完成逐页归因，以缓解跨页上下文污染和串行开销。作者同时构建了带页面级与区域级标注的跨页问答数据集 AttriDoc，并对 leader 与子代理进行角色特定训练。在三个开源骨干模型和三个数据集上，答案准确率、页面级 F1 和区域级 IoU 相比免训练版本平均提升 18.94、29.34 和 23.04 个百分点。
正文节选
AttriSwarm: Orchestrating Agent Swarms for Cross-Page Visual Attribution Abstract Multimodal large language models (MLLMs) increasingly address knowledge-intensive tasks by integrating evidence scattered across multiple document pages into a final answer. Because such answers are synthesized from many pages, the link between the answer and its supporting visual evidence can become obscured, hindering verification and trust. Cross-page visual attribution is therefore essential: it requires identi