Generalizing Graph Foundation Models via Hyperbolic Retrieval-Augmented Generation

2026-06-02

AI Summary

This work addresses the limited generalization of graph foundation models under distribution shifts and the inadequacy of existing Euclidean retrieval-augmented approaches in preserving the hierarchical semantic structure of external knowledge bases. To this end, it introduces hyperbolic geometry into the retrieval-augmented generation (RAG) framework for graph foundation models, proposing a hyperbolic knowledge index that aligns with the exponential growth property of tree-structured knowledge. The method further incorporates a multi-granularity retrieval strategy and a dual-path feature-structure fusion mechanism to mitigate semantic granularity loss and central-node bias. Experimental results demonstrate that the proposed approach significantly improves zero-shot reasoning performance across multiple graph benchmarks, validating its effectiveness in enhancing the generalization capability of graph foundation models.

Abstract

Graph foundation models (GFMs) emerged as a dominant paradigm in graph representation learning by leveraging large-scale pre-training for cross-domain inference. However, the parameterized knowledge encoded within these models is insufficient to cope with distribution shifts, limiting their generalization ability. To mitigate this issue, retrieval-augmented generation (RAG) has been introduced to incorporate external knowledge at inference time. Nevertheless, existing RAG frameworks operating in Euclidean space suffer from a fundamental geometric limitation: the polynomial volume growth of Euclidean space is inherently mismatched with the tree-structured external knowledge bases. This mismatch leads to the loss of semantic granularity in retrieval and gives rise to the hubness phenomenon.To address this limitation, we propose a Hyperbolic Retrieval-Augmented Generation (HyRAG) framework designed to enhance the generalization capabilities of GFMs. Specifically, the introduced Hyperbolic Knowledge Indexing module retains the tree-like hierarchies of the external knowledge base by modeling them within hyperbolic space. The Multi-granularity Retrieval module then provides GFMs with the global semantic anchors and local semantic nuances through coarse-grained and fine-grained knowledge retrieval, respectively. Finally, the Dual-path Fusion module achieves effective knowledge integration for graph tasks at both the feature and structural levels.Experiments on multiple graph benchmarks demonstrate significant improvements in the zero-shot setting, highlighting the generalization of our method for robust GFMs inference.

Problem

• Graph Foundation Models
• Retrieval-Augmented Generation
• Hyperbolic Space
• Distribution Shift
• Hubness Phenomenon

Innovation

• Hyperbolic Geometry
• Retrieval-Augmented Generation
• Graph Foundation Models
• Knowledge Indexing
• Zero-shot Generalization