HopRAG: Multi-Hop Reasoning for Logic-Aware Retrieval-Augmented Generation

TL;DR

HopRAG introduces a graph-based retrieval framework that enhances multi-hop reasoning and logical relevance in retrieval-augmented generation, leading to improved answer quality on multi-hop benchmarks.

Contribution

It proposes a novel retrieve-reason-prune mechanism using graph-structured knowledge and LLM-guided logical exploration for better retrieval in RAG systems.

Findings

Improves retrieval relevance through logical reasoning.

Enhances answer quality on multi-hop benchmarks.

Demonstrates effectiveness of graph-based retrieval approach.

Abstract

Retrieval-Augmented Generation (RAG) systems often struggle with imperfect retrieval, as traditional retrievers focus on lexical or semantic similarity rather than logical relevance. To address this, we propose \textbf{HopRAG}, a novel RAG framework that augments retrieval with logical reasoning through graph-structured knowledge exploration. During indexing, HopRAG constructs a passage graph, with text chunks as vertices and logical connections established via LLM-generated pseudo-queries as edges. During retrieval, it employs a \textit{retrieve-reason-prune} mechanism: starting with lexically or semantically similar passages, the system explores multi-hop neighbors guided by pseudo-queries and LLM reasoning to identify truly relevant ones. Experiments on multiple multi-hop benchmarks demonstrate that HopRAG's \textit{retrieve-reason-prune} mechanism can expand the retrieval scope based on logical connections and improve final answer quality.