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

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Poliakov, Mykhailo, Shvai, Nadiya
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912304517349376
author Poliakov, Mykhailo
Shvai, Nadiya
author_facet Poliakov, Mykhailo
Shvai, Nadiya
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2406_13213
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Multi-Meta-RAG: Improving RAG for Multi-Hop Queries using Database Filtering with LLM-Extracted Metadata
Poliakov, Mykhailo
Shvai, Nadiya
Computation and Language
Artificial Intelligence
Databases
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.
title Multi-Meta-RAG: Improving RAG for Multi-Hop Queries using Database Filtering with LLM-Extracted Metadata
topic Computation and Language
Artificial Intelligence
Databases
url https://arxiv.org/abs/2406.13213