pathGCN: Learning General Graph Spatial Operators from Paths
This addresses the problem of limited expressiveness in GCNs for researchers and practitioners in graph machine learning, offering a novel approach to improve performance.
The paper tackles the limitation of Graph Convolutional Networks (GCNs) by proposing pathGCN, a method to learn spatial operators from random paths on graphs, which inherently avoids issues like over-smoothing and achieves new state-of-the-art performance on numerous datasets.
Graph Convolutional Networks (GCNs), similarly to Convolutional Neural Networks (CNNs), are typically based on two main operations - spatial and point-wise convolutions. In the context of GCNs, differently from CNNs, a pre-determined spatial operator based on the graph Laplacian is often chosen, allowing only the point-wise operations to be learnt. However, learning a meaningful spatial operator is critical for developing more expressive GCNs for improved performance. In this paper we propose pathGCN, a novel approach to learn the spatial operator from random paths on the graph. We analyze the convergence of our method and its difference from existing GCNs. Furthermore, we discuss several options of combining our learnt spatial operator with point-wise convolutions. Our extensive experiments on numerous datasets suggest that by properly learning both the spatial and point-wise convolutions, phenomena like over-smoothing can be inherently avoided, and new state-of-the-art performance is achieved.