Combining Stochastic Explainers and Subgraph Neural Networks can Increase Expressivity and Interpretability

Authors: Indro Spinelli , Michele Guerra , Simone Scardapane , Filippo Maria Bianchi

Subgraph-enhanced GNN architecture

We designed a subgraph-enhanced GNN, which learns to sample meaningful subgraphs that work both as explanations and to increase the expressive power of the model.