DSLR: Document Refinement with Sentence-Level Re-ranking and Reconstruction to Enhance Retrieval-Augmented Generation

TL;DR

DSLR is an unsupervised framework that refines retrieved documents by sentence-level re-ranking and reconstruction, significantly improving retrieval-augmented generation performance without extra training.

Contribution

It introduces a novel, training-free method for filtering and reconstructing retrieved documents to enhance RAG systems' accuracy and relevance.

Findings

DSLR outperforms conventional fixed-size passage methods in multiple QA datasets.

It improves RAG performance without additional training.

Effective in realistic scenarios with retrieval failures.

Abstract

Recent advancements in Large Language Models (LLMs) have significantly improved their performance across various Natural Language Processing (NLP) tasks. However, LLMs still struggle with generating non-factual responses due to limitations in their parametric memory. Retrieval-Augmented Generation (RAG) systems address this issue by incorporating external knowledge with a retrieval module. Despite their successes, however, current RAG systems face challenges with retrieval failures and the limited ability of LLMs to filter out irrelevant information. Therefore, in this work, we propose DSLR (Document Refinement with Sentence-Level Re-ranking and Reconstruction), an unsupervised framework that decomposes retrieved documents into sentences, filters out irrelevant sentences, and reconstructs them again into coherent passages. We experimentally validate DSLR on multiple open-domain QA datasets and the results demonstrate that DSLR significantly enhances the RAG performance over conventional fixed-size passage. Furthermore, our DSLR enhances performance in specific, yet realistic scenarios without the need for additional training, providing an effective and efficient solution for refining retrieved documents in RAG systems.