HyperGraphRAG: Retrieval-Augmented Generation via Hypergraph-Structured Knowledge Representation

Guanting Chen (The University of North Carolina at Chapel Hill) · Anh Tuan Luu (Nanyang Technological University, Singapore) · Haoran Luo (Nanyang Technological University) · Haihong E (Beijing University of Post and Telecommunication) · Yandan Zheng (Nanyang Technological University) · Xiaobao Wu (Nanyang Technological University) · Yikai Guo (Beijing Institute of Computer Technology and Application) · Qika Lin (National University of Singapore) · Yu Feng (Shanghai Jiaotong University) · Zemin Kuang (Capital Medical University) · Meina Song (Beijing University of Posts and Telecommunications) · Yifan Zhu (Beijing University of Posts and Telecommunications)

answer accuracybinary relationscomprehensive pipelinegeneration qualitygraph-based knowledge representationgraph-based raghyperedgeshypergraph-based methodshypergraphragknowledge hypergraph constructionn-ary relationsreal-world knowledgerelational factsretrieval efficiencyretrieval-augmented generation

Standard Retrieval-Augmented Generation (RAG) relies on chunk-based retrieval, whereas GraphRAG advances this approach by graph-based knowledge representation. However, existing graph-based RAG approaches are constrained by binary relations, as each edge in an ordinary graph connects only two entities, limiting their ability to represent the n-ary relations (n >= 2) in real-world knowledge. In this work, we propose HyperGraphRAG, the first hypergraph-based RAG method that represents n-ary relational facts via hyperedges. HyperGraphRAG consists of a comprehensive pipeline, including knowledge hypergraph construction, retrieval, and generation. Experiments across medicine, agriculture, computer science, and law demonstrate that HyperGraphRAG outperforms both standard RAG and previous graph-based RAG methods in answer accuracy, retrieval efficiency, and generation quality.