Search-R2: Enhancing Search-Integrated Reasoning via Actor-Refiner Collaboration

TL;DR

Search-R2 introduces an Actor-Refiner framework for search-integrated reasoning, improving training efficiency and reasoning accuracy by targeted intervention and hybrid rewards, outperforming existing baselines.

Contribution

It presents a novel Actor-Refiner collaboration method with a dense reward mechanism, formal analysis, and extensive empirical validation for enhanced reasoning in language agents.

Findings

Outperforms strong RAG and RL baselines across datasets

Achieves higher reasoning accuracy with minimal overhead

Demonstrates theoretical performance gains through formal analysis

Abstract

Search-integrated reasoning enables language agents to transcend static parametric knowledge by actively querying external sources. However, training these agents via reinforcement learning is hindered by the multi-scale credit assignment problem: existing methods typically rely on sparse, trajectory-level rewards that fail to distinguish between high-quality reasoning and fortuitous guesses, leading to redundant or misleading search behaviors. To address this, we propose Search-R2, a novel Actor-Refiner collaboration framework that enhances reasoning through targeted intervention, with both components jointly optimized during training. Our approach decomposes the generation process into an Actor, which produces initial reasoning trajectories, and a Meta-Refiner, which selectively diagnoses and repairs flawed steps via a 'cut-and-regenerate' mechanism. To provide fine-grained supervision, we introduce a hybrid reward design that couples outcome correctness with a dense process reward quantifying the information density of retrieved evidence. Theoretically, we formalize the Actor-Refiner interaction as a smoothed mixture policy, proving that selective correction yields strict performance gains over strong baselines. Extensive experiments across various general and multi-hop QA datasets demonstrate that Search-R2 consistently outperforms strong RAG and RL-based baselines across model scales, achieving superior reasoning accuracy with minimal overhead.