GRAIN: Multi-Granular and Implicit Information Aggregation Graph Neural Network for Heterophilous Graphs

2025-04-09

AI Summary

In heterogeneous graphs, conventional graph neural networks (GNNs) underperform relative to multilayer perceptrons (MLPs) due to feature-label inconsistency and the breakdown of the homophily assumption. Method: This paper proposes a novel framework that jointly models multi-granularity explicit information aggregation and implicit long-range relation discovery—marking the first such integration. It introduces an adaptive graph information aggregator that unifies local/global structural signals with both explicit adjacency and implicit non-adjacency cues, augmented by a learnable multi-granularity weighted fusion mechanism. Contribution/Results: Evaluated on 13 diverse benchmarks spanning varying degrees of homophily, the method consistently outperforms 12 state-of-the-art approaches. It demonstrates strong generalization across both homogeneous and heterogeneous graphs, empirically validating the effectiveness and universality of explicit multi-granularity modeling coupled with implicit long-range relational characterization.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community
• Reasoning under Uncertainty: Graphical Models

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
• Web Mining and Content Analysis: Content-based information diffusion

Abstract

Graph neural networks (GNNs) have shown significant success in learning graph representations. However, recent studies reveal that GNNs often fail to outperform simple MLPs on heterophilous graph tasks, where connected nodes may differ in features or labels, challenging the homophily assumption. Existing methods addressing this issue often overlook the importance of information granularity and rarely consider implicit relationships between distant nodes. To overcome these limitations, we propose the Granular and Implicit Graph Network (GRAIN), a novel GNN model specifically designed for heterophilous graphs. GRAIN enhances node embeddings by aggregating multi-view information at various granularity levels and incorporating implicit data from distant, non-neighboring nodes. This approach effectively integrates local and global information, resulting in smoother, more accurate node representations. We also introduce an adaptive graph information aggregator that efficiently combines multi-granularity and implicit data, significantly improving node representation quality, as shown by experiments on 13 datasets covering varying homophily and heterophily. GRAIN consistently outperforms 12 state-of-the-art models, excelling on both homophilous and heterophilous graphs.

Problem

• Improving GNN performance on heterophilous graphs
• Aggregating multi-granular information for better node embeddings
• Incorporating implicit relationships from distant nodes

Innovation

• Multi-granular information aggregation for heterophilous graphs
• Implicit data integration from distant nodes
• Adaptive graph information aggregator