From Ranking to Selection: A Simple but Efficient Dynamic Passage Selector for Retrieval Augmented Generation
This addresses a key limitation in RAG systems for complex queries, though it appears incremental as a novel method for a known bottleneck.
The paper tackles the bottleneck in retrieval-augmented generation (RAG) systems where fixed Top-K passage selection struggles with complex multi-hop queries, introducing the Dynamic Passage Selector (DPS) that dynamically selects relevant passages. On the MuSiQue dataset, DPS improves F1-score by 30.06% and 15.4% over strong baselines like Qwen3-reranker and RankingGPT.
Retrieval-augmented generation (RAG) systems are often bottlenecked by their reranking modules, which typically score passages independently and select a fixed Top-K size. This approach struggles with complex multi-hop queries that require synthesizing evidence across multiple documents, creating a trade-off where small K values omit crucial information and large K values introduce noise. To address this, we introduce the Dynamic Passage Selector (DPS), a novel reranking framework that treats passage selection as a supervised learning problem. Unlike traditional point-wise or list-wise methods, DPS is fine-tuned to capture inter-passage dependencies and dynamically select the most relevant set of passages for generation. As a seamless plug-and-play module, DPS requires no modifications to the standard RAG pipeline. Comprehensive evaluations on five benchmarks show that DPS consistently outperforms state-of-the-art rerankers and fine-tuning methods. Notably, on the challenging MuSiQue dataset, DPS improves the F1-score by 30.06% and 15.4% over strong baselines like Qwen3-reranker and RankingGPT, respectively. Our results demonstrate that by enabling adaptive evidence selection, DPS substantially enhances reasoning capabilities in complex RAG scenarios.