Revisiting Heterophily in Graph Convolution Networks by Learning Representations Across Topological and Feature Spaces

TL;DR

This paper proposes a new GCN framework that learns and combines representations across topological and feature spaces to improve performance on both homophilous and heterophilous graphs.

Contribution

It introduces a novel approach of learning across two spaces, addressing heterophily in GCNs, which was not effectively handled by previous methods.

Findings

Improved semi-supervised node classification on heterophilous graphs.

Effective representation learning across topology and feature spaces.

Outperforms existing methods on benchmark datasets.

Abstract

Graph convolution networks (GCNs) have been enormously successful in learning representations over several graph-based machine learning tasks. Specific to learning rich node representations, most of the methods have solely relied on the homophily assumption and have shown limited performance on the heterophilous graphs. While several methods have been developed with new architectures to address heterophily, we argue that by learning graph representations across two spaces i.e., topology and feature space GCNs can address heterophily. In this work, we experimentally demonstrate the performance of the proposed GCN framework over semi-supervised node classification task on both homophilous and heterophilous graph benchmarks by learning and combining representations across the topological and the feature spaces.