Rethinking Explaining Graph Neural Networks via Non-parametric Subgraph Matching

Open in new window