Adaptive Heterogeneous Graph Neural Networks: Bridging Heterophily and Heterogeneity

2025-08-08

AI Summary

Existing methods typically model heterogeneity or heterophily of heterogeneous graphs in isolation, failing to adequately address the prevalent yet challenging heterophilous heterogeneous graphs. To bridge this gap, we propose Adaptive Heterogeneous Graph Neural Networks (AHGNN), the first framework that jointly models heterophily distributions across hop distances and meta-paths. AHGNN introduces fine-grained heterophily-aware graph convolution and multi-semantic-space noise filtering. Moreover, it incorporates meta-path-based semantic decomposition and coarse-to-fine heterogeneous graph attention to jointly capture structural and semantic information. Extensive experiments on seven real-world datasets against twenty baselines demonstrate that AHGNN significantly outperforms state-of-the-art methods on node classification—particularly under high-heterophily regimes—where it exhibits superior robustness and generalization.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community
• Search and Optimization: Metareasoning and Metaheuristics

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Social Networks and Social Media: Influence propagation, information diffusion, and the prediction on networks

Abstract

Heterogeneous graphs (HGs) are common in real-world scenarios and often exhibit heterophily. However, most existing studies focus on either heterogeneity or heterophily in isolation, overlooking the prevalence of heterophilic HGs in practical applications. Such ignorance leads to their performance degradation. In this work, we first identify two main challenges in modeling heterophily HGs: (1) varying heterophily distributions across hops and meta-paths; (2) the intricate and often heterophily-driven diversity of semantic information across different meta-paths. Then, we propose the Adaptive Heterogeneous Graph Neural Network (AHGNN) to tackle these challenges. AHGNN employs a heterophily-aware convolution that accounts for heterophily distributions specific to both hops and meta-paths. It then integrates messages from diverse semantic spaces using a coarse-to-fine attention mechanism, which filters out noise and emphasizes informative signals. Experiments on seven real-world graphs and twenty baselines demonstrate the superior performance of AHGNN, particularly in high-heterophily situations.

Problem

• Modeling heterophily distributions in heterogeneous graphs
• Handling semantic diversity across different meta-paths
• Improving performance in high-heterophily graph scenarios

Innovation

• Heterophily-aware convolution for HGNN
• Coarse-to-fine attention mechanism
• Adaptive handling of heterophily distributions