让检索模型理解研究代理的思考过程，显著提升搜索准确率
AgentIR: Reasoning-Aware Retrieval for Deep Research Agents
- 将代理的推理过程与查询一同嵌入，捕捉深层意图
- 在BrowseComp-Plus上达68%准确率，远超传统方法
- 适合构建智能研究代理的研究者和开发者
深度研究代理正成为现代检索系统的主要使用者。与人类用户不同，这些代理在每次搜索前会生成明确的自然语言推理过程，暴露丰富的意图和上下文信息，但现有检索器完全忽视这一信号。为此，我们提出：(1) 推理感知检索，联合嵌入代理的推理轨迹与查询；(2) DR-Synth，一种从标准问答数据集合成深度研究检索训练数据的方法。我们证明两者独立有效，其结合可训练出性能显著提升的嵌入模型AgentIR-4B。在挑战性任务BrowseComp-Plus上，AgentIR-4B使用开源代理Tongyi-DeepResearch时达到68%准确率，远超两倍大小的传统模型（50%）和BM25（37%）。代码与数据见：https://texttron.github.io/AgentIR/
原文摘要 · Abstract (English)
Deep Research agents are rapidly emerging as primary consumers of modern retrieval systems. Unlike human users who issue and refine queries without documenting their intermediate thought processes, Deep Research agents generate explicit natural language reasoning before each search call, revealing rich intent and contextual information that existing retrievers entirely ignore. To exploit this overlooked signal, we introduce: (1) Reasoning-Aware Retrieval, a retrieval paradigm that jointly embeds the agent's reasoning trace alongside its query; and (2) DR-Synth, a data synthesis method that generates Deep Research retriever training data from standard QA datasets. We demonstrate that both components are independently effective, and their combination yields a trained embedding model, AgentIR-4B, with substantial gains. On the challenging BrowseComp-Plus benchmark, AgentIR-4B achieves 68\% accuracy with the open-weight agent Tongyi-DeepResearch, compared to 50\% with conventional embedding models twice its size, and 37\% with BM25. Code and data are available at: https://texttron.github.io/AgentIR/.
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。