Learn from Heterophily: Heterophilous Information-enhanced Graph Neural Network

TL;DR

This paper introduces HiGNN, a novel graph neural network that leverages heterophilous information by constructing an additional graph structure based on node distribution, thereby improving performance on heterophilous and homophilous datasets.

Contribution

The paper proposes HiGNN, which effectively utilizes heterophilous semantic information through a new graph structure, demonstrating theoretical and empirical improvements over existing GNN methods.

Findings

HiGNN outperforms baseline GNNs on heterophilous datasets.

Incorporating heterophilous information enhances existing GNN approaches.

The method improves homophily degree in real-world datasets.

Abstract

Under circumstances of heterophily, where nodes with different labels tend to be connected based on semantic meanings, Graph Neural Networks (GNNs) often exhibit suboptimal performance. Current studies on graph heterophily mainly focus on aggregation calibration or neighbor extension and address the heterophily issue by utilizing node features or structural information to improve GNN representations. In this paper, we propose and demonstrate that the valuable semantic information inherent in heterophily can be utilized effectively in graph learning by investigating the distribution of neighbors for each individual node within the graph. The theoretical analysis is carried out to demonstrate the efficacy of the idea in enhancing graph learning. Based on this analysis, we propose HiGNN, an innovative approach that constructs an additional new graph structure, that integrates heterophilous information by leveraging node distribution to enhance connectivity between nodes that share similar semantic characteristics. We conduct empirical assessments on node classification tasks using both homophilous and heterophilous benchmark datasets and compare HiGNN to popular GNN baselines and SoTA methods, confirming the effectiveness in improving graph representations. In addition, by incorporating heterophilous information, we demonstrate a notable enhancement in existing GNN-based approaches, and the homophily degree across real-world datasets, thus affirming the efficacy of our approach.