检索感知代理控制器缓解深度研究代理的推理停滞
原标题：When Deep Research Agents Stagnate: Enhancing Reasoning with Retrieval-Aware Agent Control
AI 摘要
本文分析了多种深度研究代理（DRA）的推理轨迹，发现它们常出现“推理停滞”问题，即多数迭代对最终性能贡献甚微，且代理缺乏对自身轨迹的感知。为此，作者提出无监督信号和检索感知代理控制器（RAAC），帮助代理在每步选择最优动作，减少不必要的搜索调用并提升性能。在BrowseComp-Plus等基准上，RAAC平均减少14次搜索调用，准确率最高提升10%，平均提升3%。
正文节选
When Deep Research Agents Stagnate: Enhancing Reasoning with Retrieval-Aware Agent Control Abstract In this paper, we analyze the reasoning trajectories of a variety of DRAs and show that existing agents often suffer from reasoning stagnation: the majority of iterations contribute little or no improvement to final performance, while agents lack awareness of their trajectories and are therefore ineffective at adapting their search strategies or determining when to terminate. To address this issue