Relevance to Utility: Process-Supervised Rewrite for RAG
This addresses a key bottleneck in RAG systems for improving answer accuracy, though it appears incremental relative to existing bridge modules.
The paper tackles the gap between retrieval relevance and generative utility in Retrieval-Augmented Generation systems by proposing R2U, which directly optimizes for correct answer generation through process supervision, showing consistent improvements over baselines on open-domain question-answering benchmarks.
Retrieval-Augmented Generation systems often suffer from a gap between optimizing retrieval relevance and generative utility: retrieved documents may be topically relevant but still lack the content needed for effective reasoning during generation. While existing "bridge" modules attempt to rewrite the retrieved text for better generation, we show how they fail to capture true document utility. In this work, we propose R2U, with a key distinction of directly optimizing to maximize the probability of generating a correct answer through process supervision. As such direct observation is expensive, we also propose approximating an efficient distillation pipeline by scaling the supervision from LLMs, which helps the smaller rewriter model generalize better. We evaluate our method across multiple open-domain question-answering benchmarks. The empirical results demonstrate consistent improvements over strong bridging baselines.