Multi-Meta-RAG: Improving RAG for Multi-Hop Queries using Database Filtering with LLM-Extracted Metadata

TL;DR

Multi-Meta-RAG enhances retrieval-augmented generation for multi-hop questions by using LLM-extracted metadata for database filtering, significantly improving performance on the MultiHop-RAG benchmark.

Contribution

Introduces Multi-Meta-RAG, a novel method that employs database filtering with LLM-extracted metadata to improve multi-hop question answering in RAG systems.

Findings

Significant performance improvement on MultiHop-RAG benchmark

Effective use of LLM-extracted metadata for document filtering

Applicable to domain-specific multi-hop question answering

Abstract

The retrieval-augmented generation (RAG) enables retrieval of relevant information from an external knowledge source and allows large language models (LLMs) to answer queries over previously unseen document collections. However, it was demonstrated that traditional RAG applications perform poorly in answering multi-hop questions, which require retrieving and reasoning over multiple elements of supporting evidence. We introduce a new method called Multi-Meta-RAG, which uses database filtering with LLM-extracted metadata to improve the RAG selection of the relevant documents from various sources, relevant to the question. While database filtering is specific to a set of questions from a particular domain and format, we found out that Multi-Meta-RAG greatly improves the results on the MultiHop-RAG benchmark. The code is available at https://github.com/mxpoliakov/Multi-Meta-RAG.