edGNN: a Simple and Powerful GNN for Directed Labeled Graphs
This work addresses inference problems on directed graphs with node and edge labels, but it appears incremental as it builds on previous work to extend capabilities to directed labeled graphs.
The authors tackled the problem of building discriminative embeddings for directed labeled graphs using graph neural networks, and they theoretically and experimentally demonstrated that their model, edGNN, is as powerful as the Weisfeiler-Lehman algorithm for graph isomorphism.
The ability of a graph neural network (GNN) to leverage both the graph topology and graph labels is fundamental to building discriminative node and graph embeddings. Building on previous work, we theoretically show that edGNN, our model for directed labeled graphs, is as powerful as the Weisfeiler-Lehman algorithm for graph isomorphism. Our experiments support our theoretical findings, confirming that graph neural networks can be used effectively for inference problems on directed graphs with both node and edge labels. Code available at https://github.com/guillaumejaume/edGNN.