Contrastive Graph Neural Network Explanation

Understand

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.