Problem
Research questions and friction points this paper is trying to address.
Graph Retrieval-Augmented Generation
Multi-hop Retrieval
Relation-free Graph Retrieval
Bridging Evidence
Entity Activation
Corpus-Guided Dual-Path Propagation for Graph Retrieval-Augmented Generation
🤖 AI Summary
This study addresses the limitation of existing graph retrieval methods, which over-rely on query similarity and consequently overlook bridging evidence while activating irrelevant entities. To this end, we propose NexusRAG, a novel framework that introduces a dual-path propagation mechanism integrating semantic and structural information to guide multi-hop reasoning via corpus-level entity neighborhoods. Built upon Tri-Graph construction and entity co-occurrence analysis, the method further optimizes the passage initialization strategy of Personalized PageRank to enhance retrieval precision. Experimental results demonstrate that NexusRAG achieves superior performance on multi-hop question answering benchmarks, significantly improving evidence recall by 4.2 to 8.1 percentage points over existing baselines.
📝 Abstract
Graph-based retrieval-augmented generation supports multi-hop retrieval by organizing corpus information into graphs. However, existing relation-free graph retrieval methods rely primarily on query-sentence similarity to search for evidence. This can exclude useful bridging evidence with low query similarity and activate incidental entities unrelated to the reasoning chain. In this paper, we propose a simple and effective approach called NexusRAG, which augments the relation-free Tri-Graph with a corpus-level entity neighborhood structure derived from joint entity co-occurrence and semantic similarity. NexusRAG employs this structure to guide two complementary propagation paths: neighborhood-constrained semantic propagation through sentences identifies the query-relevant entity frontier, while direct structural propagation between neighboring entities expands that frontier to structurally related entities. The propagated entity weights also inform neighborhood-aware passage initialization for Personalized PageRank. Experiments on three multi-hop QA benchmarks and a domain-specific subset of GraphRAG-Bench show that NexusRAG consistently outperforms existing approaches. On the GraphRAG-Bench subset, NexusRAG achieves the highest evidence recall in all question categories, exceeding baselines by 4.2-8.1 points. The implementation code is available at https://github.com/Jacob-biu/NexusRAG.
Innovation
Methods, ideas, or system contributions that make the work stand out.
Graph Retrieval-Augmented Generation
Dual-Path Propagation
Entity Neighborhood Structure
Personalized PageRank
Multi-hop QA
🔎 Similar Papers
No similar papers found.
