DF-RAG: Query-Aware Diversity for Retrieval-Augmented Generation

TL;DR

DF-RAG enhances retrieval-augmented generation by dynamically incorporating diversity into retrieval, significantly improving performance on complex reasoning questions without extra training.

Contribution

Introduces a novel method that optimizes diversity in retrieval for RAG, improving reasoning-intensive QA performance without additional fine-tuning.

Findings

DF-RAG improves F1 scores by 4-10% over vanilla RAG.

DF-RAG captures up to 91.3% of the estimated Oracle ceiling.

Outperforms established baselines on reasoning-intensive benchmarks.

Abstract

Retrieval-augmented generation (RAG) is a common technique for grounding language model outputs in domain-specific information. However, RAG is often challenged by reasoning-intensive question-answering (QA), since common retrieval methods like cosine similarity maximize relevance at the cost of introducing redundant content, which can reduce information recall. To address this, we introduce Diversity-Focused Retrieval-Augmented Generation (DF-RAG), which systematically incorporates diversity into the retrieval step to improve performance on complex, reasoning-intensive QA benchmarks. DF-RAG builds upon the Maximal Marginal Relevance framework to select information chunks that are both relevant to the query and maximally dissimilar from each other. A key innovation of DF-RAG is its ability to optimize the level of diversity for each query dynamically at test time without requiring any additional fine-tuning or prior information. We show that DF-RAG improves F1 performance on reasoning-intensive QA benchmarks by 4-10 percent over vanilla RAG using cosine similarity and also outperforms other established baselines. Furthermore, we estimate an Oracle ceiling of up to 18 percent absolute F1 gains over vanilla RAG, of which DF-RAG captures up to 91.3 percent.