Shifting from Ranking to Set Selection for Retrieval Augmented Generation
By: Dahyun Lee , Yongrae Jo , Haeju Park and more
Potential Business Impact:
Finds better answers by picking groups of facts.
Retrieval in Retrieval-Augmented Generation(RAG) must ensure that retrieved passages are not only individually relevant but also collectively form a comprehensive set. Existing approaches primarily rerank top-k passages based on their individual relevance, often failing to meet the information needs of complex queries in multi-hop question answering. In this work, we propose a set-wise passage selection approach and introduce SETR, which explicitly identifies the information requirements of a query through Chain-of-Thought reasoning and selects an optimal set of passages that collectively satisfy those requirements. Experiments on multi-hop RAG benchmarks show that SETR outperforms both proprietary LLM-based rerankers and open-source baselines in terms of answer correctness and retrieval quality, providing an effective and efficient alternative to traditional rerankers in RAG systems. The code is available at https://github.com/LGAI-Research/SetR
Similar Papers
OptiSet: Unified Optimizing Set Selection and Ranking for Retrieval-Augmented Generation
Artificial Intelligence
Finds better information for smarter computer answers.
LTRR: Learning To Rank Retrievers for LLMs
Computation and Language
Chooses best search tool for better answers.
Chain-of-Retrieval Augmented Generation
Information Retrieval
Helps computers answer harder questions by thinking more.