Adaptive Heterogeneous Graph Neural Networks: Bridging Heterophily and Heterogeneity
This addresses the performance degradation in graph neural networks for real-world applications like social or biological networks where heterophily and heterogeneity coexist, representing an incremental improvement over prior work.
The paper tackled the problem of modeling heterophilic heterogeneous graphs, where existing methods overlook the interplay of heterophily and heterogeneity, by proposing AHGNN, which achieved superior performance on seven real-world graphs, especially in high-heterophily scenarios.
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.