Beyond Feature and Structure Alignment: Learning Transferable Propagation Knowledge for Graph Foundation Models

2026-07-30

AI Summary

Existing graph foundation models lack a unified, transferable knowledge unit, as they align only node features and structural properties while overlooking the heterogeneity of propagation patterns, thereby limiting their generalization to unseen graph domains. To address this, this work proposes Propagation-aware Graph Foundation Model (ProGFM), which, for the first time, formulates the propagation relationships between edges and feature dimensions as a transferable knowledge unit. ProGFM constructs a prototype library of propagation relations to capture heterogeneous propagation patterns across domains, enabling adaptive message passing and cross-domain graph representation learning. Experimental results demonstrate that ProGFM significantly outperforms current state-of-the-art models across diverse cross-domain transfer scenarios, exhibiting superior generalization and knowledge transfer capabilities.

Abstract

Graph Foundation Models (GFMs) have recently emerged as a promising paradigm for enabling knowledge transfer across diverse domains. Unlike traditional graph learning methods that are typically designed for in-domain settings, GFMs aim to learn transferable knowledge that can generalize to unseen graph domains. However, unlike language or visual data, graphs lack intrinsic and unified representation units, such as tokens in language and patches in vision, making it challenging to identify transferable knowledge units for building graph foundation models. Existing graph foundation models mainly focus on mitigating domain discrepancies through feature alignment and structure alignment, while overlooking the exploration of transferable knowledge units underlying graph data. Moreover, these methods generally rely on fixed propagation mechanisms during message passing, overlooking the heterogeneity in propagation patterns, as different edges may exhibit distinct propagation patterns for different feature dimensions. To address these limitations, we propose a Propagation-aware Graph Foundation Model (ProGFM), which regards the propagation relationships between edges and feature dimensions as transferable knowledge units. Through a propagation relationship prototype bank, ProGFM learns cross-domain transferable propagation knowledge, enabling adaptive information aggregation in unseen graph domains. Extensive experiments across various cross-domain transfer scenarios demonstrate that ProGFM possesses strong cross-domain knowledge transfer capability and exhibits superior generalization performance compared with existing methods.

Problem

• Graph Foundation Models
• Transferable Knowledge
• Propagation Patterns
• Cross-domain Generalization
• Message Passing

Innovation

• Propagation-aware
• Graph Foundation Models
• Transferable Knowledge
• Cross-domain Generalization
• Message Passing