Graph Neural Networks with Heterophily
This addresses a limitation in GNNs for graph learning tasks where heterophily is common, such as in social or biological networks, representing a novel method rather than an incremental improvement.
The authors tackled the problem that many Graph Neural Networks (GNNs) assume homophily, overlooking heterophily where connected nodes are from different classes, by proposing CPGNN, a framework that generalizes GNNs for both homophily and heterophily graphs, achieving state-of-the-art results in heterophily settings with significantly less training data.
Graph Neural Networks (GNNs) have proven to be useful for many different practical applications. However, many existing GNN models have implicitly assumed homophily among the nodes connected in the graph, and therefore have largely overlooked the important setting of heterophily, where most connected nodes are from different classes. In this work, we propose a novel framework called CPGNN that generalizes GNNs for graphs with either homophily or heterophily. The proposed framework incorporates an interpretable compatibility matrix for modeling the heterophily or homophily level in the graph, which can be learned in an end-to-end fashion, enabling it to go beyond the assumption of strong homophily. Theoretically, we show that replacing the compatibility matrix in our framework with the identity (which represents pure homophily) reduces to GCN. Our extensive experiments demonstrate the effectiveness of our approach in more realistic and challenging experimental settings with significantly less training data compared to previous works: CPGNN variants achieve state-of-the-art results in heterophily settings with or without contextual node features, while maintaining comparable performance in homophily settings.