GNNExplainer: Generating Explanations for Graph Neural Networks

Authors: Rex Ying, Dylan Bourgeois, Jiaxuan You, Marinka Zitnik, Jure Leskovec
Year: 2019
Summary: This paper introduces GNNExplainer, a method for explaining the predictions of Graph Neural Networks (GNNs). It identifies a compact subgraph and a small subset of node features that are most influential for a given prediction. This is crucial for understanding GNN decisions in domains like social networks or knowledge graphs.
Link: https://arxiv.org/abs/1903.03894