Directed Homophily-Aware Graph Neural Network

2025/05/28 by Aihu Zhang, Zhang, Aihu, Jiaxing Xu +7
Computer Science · Physics and Astronomy · #Advanced Graph Neural Networks #Complex Network Analysis Techniques #FOS: Computer and information sciences #Graph Theory and Algorithms #Machine Learning (cs.LG)

paper · pdf · doi:10.48550/arxiv.2505.22362

openalex publication_date 2025/05/28 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Graph Neural Networks (GNNs) have achieved significant success in various learning tasks on graph-structured data. Nevertheless, most GNNs struggle to generalize to heterophilic neighborhoods. Additionally, many GNNs ignore the directional nature of real-world graphs, resulting in suboptimal performance on directed graphs with asymmetric structures. In this work, we propose Directed Homophily-aware Graph Neural Network (DHGNN), a novel framework that addresses these limitations by incorporating homophily-aware and direction-sensitive components. DHGNN employs a resettable gating mechanism to adaptively modulate message contributions based on homophily levels and informativeness, and a structure-aware noise-tolerant fusion module to effectively integrate node representations from the original and reverse directions. Extensive experiments on both homophilic and heterophilic directed graph datasets demonstrate that DHGNN outperforms state-of-the-art methods in node classification and link prediction. In particular, DHGNN improves over the best baseline by up to 15.07% in link prediction. Our analysis further shows that the gating mechanism captures directional homophily gaps and fluctuating homophily across layers, providing deeper insights into message-passing behavior on complex graph structures.

Citations

• What Are Good Positional Encodings for Directed Graphs?
• Robust Graph Structure Learning under Heterophily
• Edge Directionality Improves Learning on Heterophilic Graphs
• Graph Mixture of Experts: Learning on Large-Scale Graphs with Explicit Diversity Modeling
• A critical look at the evaluation of GNNs under heterophily: Are we really making progress?
• Ordered GNN: Ordering Message Passing to Deal with Heterophily and Over-smoothing
• Transformers Meet Directed Graphs
• Revisiting Heterophily For Graph Neural Networks
• Gradient Gating for Deep Multi-Rate Learning on Graphs
• Finding Global Homophily in Graph Neural Networks When Meeting Heterophily
• Block Modeling-Guided Graph Convolutional Neural Networks
• Large Scale Learning on Non-Homophilous Graphs: New Benchmarks and Strong Simple Methods
• Improving Graph Neural Networks with Simple Architecture Design
• MagNet: A Neural Network for Directed Graphs
• Two Sides of the Same Coin: Heterophily and Oversmoothing in Graph Convolutional Neural Networks
• Adaptive Universal Generalized PageRank Graph Neural Network
• Directed Graph Convolutional Network
• Geom-GCN: Geometric Graph Convolutional Networks
• Optuna: A Next-generation Hyperparameter Optimization Framework
• Fast Graph Representation Learning with PyTorch Geometric
• How Powerful are Graph Neural Networks?
• Learning Word Vectors for 157 Languages
• Deeper Insights into Graph Convolutional Networks for Semi-Supervised Learning
• Graph Attention Networks
• Inductive Representation Learning on Large Graphs
• Neural Message Passing for Quantum Chemistry
• Semi-Supervised Classification with Graph Convolutional Networks
• The Graph Neural Network Model

Related

• Edge Directionality Improves Learning on Heterophilic Graphs
• Is Homophily a Necessity for Graph Neural Networks?
• Adapting to Heterophilic Graph Data with Structure-Guided Neighbor Discovery
• Flow Matters: Directional and Expressive GNNs for Heterophilic Graphs
• Graph Neural Networks for Graphs with Heterophily: A Survey