Optimizing RAG Rerankers with LLM Feedback via Reinforcement Learning

TL;DR

This paper presents RRPO, a reinforcement learning framework that directly optimizes rerankers for LLM generation quality, reducing reliance on human relevance labels and improving retrieval-augmented generation performance.

Contribution

It introduces a novel RL-based reranking approach aligned with LLM utility, enhancing retrieval results for better downstream generation without extensive human annotations.

Findings

RRPO outperforms existing rerankers like RankZephyr on knowledge-intensive benchmarks.

The framework generalizes to different LLMs such as GPT-4o.

It remains robust even with noisy supervision signals.

Abstract

Rerankers play a pivotal role in refining retrieval results for Retrieval-Augmented Generation. However, current reranking models are typically optimized on static human annotated relevance labels in isolation, decoupled from the downstream generation process. This isolation leads to a fundamental misalignment: documents identified as topically relevant by information retrieval metrics often fail to provide the actual utility required by the LLM for precise answer generation. To bridge this gap, we introduce ReRanking Preference Optimization (RRPO), a reinforcement learning framework that directly aligns reranking with the LLM's generation quality. By formulating reranking as a sequential decision-making process, RRPO optimizes for context utility using LLM feedback, thereby eliminating the need for expensive human annotations. To ensure training stability, we further introduce a reference-anchored deterministic baseline. Extensive experiments on knowledge-intensive benchmarks demonstrate that RRPO significantly outperforms strong baselines, including the powerful list-wise reranker RankZephyr. Further analysis highlights the versatility of our framework: it generalizes seamlessly to diverse readers (e.g., GPT-4o), integrates orthogonally with query expansion modules like Query2Doc, and remains robust even when trained with noisy supervisors.