Spectral Multigraph Networks for Discovering and Fusing Relationships in Molecules
This work addresses graph classification tasks with variable structures for applications in chemistry, but it is incremental as it builds on existing Chebyshev GCNs.
The authors tackled the limitation of spectral graph convolutional networks to fixed graphs by proposing a novel multigraph network for learning from multi-relational graphs, achieving competitive results on chemical classification benchmarks.
Spectral Graph Convolutional Networks (GCNs) are a generalization of convolutional networks to learning on graph-structured data. Applications of spectral GCNs have been successful, but limited to a few problems where the graph is fixed, such as shape correspondence and node classification. In this work, we address this limitation by revisiting a particular family of spectral graph networks, Chebyshev GCNs, showing its efficacy in solving graph classification tasks with a variable graph structure and size. Chebyshev GCNs restrict graphs to have at most one edge between any pair of nodes. To this end, we propose a novel multigraph network that learns from multi-relational graphs. We model learned edges with abstract meaning and experiment with different ways to fuse the representations extracted from annotated and learned edges, achieving competitive results on a variety of chemical classification benchmarks.