Contrastive Graph Neural Network Explanation
This addresses the need for reliable explanations in GNNs for users in graph-based machine learning, though it appears incremental as it builds on existing explanation challenges.
The paper tackles the problem of explaining Graph Neural Networks (GNNs) by proposing a Distribution Compliant Explanation (DCE) paradigm to avoid model confusion from out-of-distribution graphs, and presents CoGE, a novel technique that shows efficacy in experiments.
Graph Neural Networks achieve remarkable results on problems with structured data but come as black-box predictors. Transferring existing explanation techniques, such as occlusion, fails as even removing a single node or edge can lead to drastic changes in the graph. The resulting graphs can differ from all training examples, causing model confusion and wrong explanations. Thus, we argue that explicability must use graphs compliant with the distribution underlying the training data. We coin this property Distribution Compliant Explanation (DCE) and present a novel Contrastive GNN Explanation (CoGE) technique following this paradigm. An experimental study supports the efficacy of CoGE.