R$^2$-Searcher: Calibrating Retrieval and Reasoning Boundaries for Agentic Search
For developers of search agents and QA systems, this work addresses a known bottleneck in multi-hop reasoning by explicitly calibrating retrieval and reasoning boundaries, though it is an incremental improvement over existing agentic search frameworks.
R^2-Searcher tackles the problem of retrieval-reasoning boundary shift in multi-hop search agents, where incomplete evidence or irrelevant reasoning degrades performance. It achieves state-of-the-art results on seven benchmarks, significantly outperforming existing methods in answer accuracy and retrieval-reasoning quality.
Recent search agents for multi-hop reasoning often fail by either retrieving incomplete evidence or reasoning over irrelevant portions of the retrieved content, leading to a retrieval-reasoning boundary shift. We propose R$^2$-Searcher, a novel framework that explicitly explores and calibrates the retrieval and reasoning boundaries via fine-grained, query-token-guided evidence modeling and post-retrieval reflection. Specifically, R$^2$-Searcher: (1) constructs fine-grained reasoning contexts by extracting precise facts from retrieved content based on query token semantics (e.g., subjects, actions, temporal markers, and degree modifiers), thereby guiding the attention of search agent; (2) introduces a retrieval reflection mechanism that evaluates and corrects boundary deviations after each retrieval step, guiding the generation of improved queries grounded in the extracted reasoning contexts; and (3) employs an end-to-end reasoning-reflection-guided reinforcement learning algorithm, R$^2$PO, which jointly optimizes both boundaries through a tree-based exploration of reasoning regions and reflections. Our method significantly enhances the quality of both retrieval and reasoning, establishing an iterative loop where retrieval and reasoning mutually enhance each other. Extensive experiments on seven complex multi-hop QA benchmarks demonstrate that R$^2$-Searcher significantly outperforms state-of-the-art agentic search methods in answer accuracy and retrieval-reasoning quality. Ablation studies further confirm the critical role of retrieval-reasoning boundary calibration.