Cog-RAG: Cognitive-Inspired Dual-Hypergraph with Theme Alignment Retrieval-Augmented Generation

2025-11-17

AI Summary

Existing RAG methods primarily model low-order pairwise entity relations, failing to capture high-order semantic associations across text chunks; while hypergraph-enhanced approaches introduce hyperedges, they often neglect global thematic structure and inter-chunk topic alignment. This paper proposes a cognitively inspired dual-hypergraph-enhanced RAG framework: it constructs a *topic hypergraph* to model cross-chunk global thematic organization and an *entity hypergraph* to encode high-order multi-entity interactions, coupled with a two-stage retrieval mechanism for topic-guided fine-grained semantic alignment in generation. The method integrates hypergraph neural networks, probabilistic topic modeling, and diffusion-based reasoning. Experiments demonstrate that our framework significantly outperforms state-of-the-art baselines across multiple benchmarks, effectively mitigating hallucination while improving response relevance, coherence, and semantic consistency.

Technology Category

• Knowledge Representation and Reasoning: Computational Complexity of Reasoning
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning
• Data Mining & Knowledge Management: Linked Open Data, Knowledge Graphs & KB Completion

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Retrieval-Augmented Generation (RAG) enhances the response quality and domain-specific performance of large language models (LLMs) by incorporating external knowledge to combat hallucinations. In recent research, graph structures have been integrated into RAG to enhance the capture of semantic relations between entities. However, it primarily focuses on low-order pairwise entity relations, limiting the high-order associations among multiple entities. Hypergraph-enhanced approaches address this limitation by modeling multi-entity interactions via hyperedges, but they are typically constrained to inter-chunk entity-level representations, overlooking the global thematic organization and alignment across chunks. Drawing inspiration from the top-down cognitive process of human reasoning, we propose a theme-aligned dual-hypergraph RAG framework (Cog-RAG) that uses a theme hypergraph to capture inter-chunk thematic structure and an entity hypergraph to model high-order semantic relations. Furthermore, we design a cognitive-inspired two-stage retrieval strategy that first activates query-relevant thematic content from the theme hypergraph, and then guides fine-grained recall and diffusion in the entity hypergraph, achieving semantic alignment and consistent generation from global themes to local details. Our extensive experiments demonstrate that Cog-RAG significantly outperforms existing state-of-the-art baseline approaches.

Problem

• Capturing high-order semantic relations among multiple entities in RAG systems
• Modeling global thematic organization across document chunks in retrieval
• Achieving semantic alignment from global themes to local details

Innovation

• Theme hypergraph captures inter-chunk thematic structure
• Entity hypergraph models high-order semantic relations
• Two-stage retrieval aligns global themes with local details