On the Impact of Feature Heterophily on Link Prediction with Graph Neural Networks

Jiong Zhu, Gaotang Li, Yao-An Yang, Jing Zhu, Xuehao Cui, Danai Koutra

Keywords: graph neural networks, heterophily, link prediction

TL;DR: We formalize and analyze heterophilic link prediction with GNNs, when connected nodes have dissimilar features. We identify real-world heterophilic benchmarks and show that learnable decoders and separated node embeddings are crucial for such graphs.

Abstract: Heterophily, or the tendency of connected nodes in networks to have different class labels or dissimilar features, has been identified as challenging for many Graph Neural Network (GNN) models. While the challenges of applying GNNs for node classification when class labels display strong heterophily are well understood, it is unclear how heterophily affects GNN performance in other important graph learning tasks where class labels are not available. In this work, we focus on the link prediction task and systematically analyze the impact of heterophily in node features on GNN performance. We first introduce formal definitions of homophilic and heterophilic link prediction tasks, and present a theoretical framework that highlights the different optimizations needed for the respective tasks. We then analyze how different link prediction encoders and decoders adapt to varying levels of feature homophily and introduce designs for improved performance. Based on our definitions, we identify and analyze six real-world benchmarks spanning from homophilic to heterophilic link prediction settings, with graphs containing up to 30M edges. Our empirical analysis on a variety of synthetic and real-world datasets confirms our theoretical insights and highlights the importance of adopting learnable decoders and GNN encoders with ego- and neighbor-embedding separation in message passing for link prediction tasks beyond homophily.

Primary Area: Graph neural networks

Submission Number: 14119