Timezone: »

Counterfactual Explanations for Graph Neural Networks
Ana Lucic · Maartje ter Hoeve · Gabriele Tolomei · Maarten de Rijke · Fabrizio Silvestri

Given the increasing promise of Graph Neural Networks (GNNs) in real-world applications, several methods have been developed for explaining their predictions. However, such methods do not provide a clear opportunity for recourse: given a prediction, we want to understand how the prediction can be changed in order to achieve a more desirable outcome. In this work, we propose a method for generating counterfactual (CF) explanations for GNNs: the minimal perturbation to the input (graph) data such that the prediction changes. Using only edge deletions, we find that our method can generate CF explanations for the majority of instances across three widely used datasets for GNN explanations, while removing less than 3 edges on average, with at least 94\% accuracy. This indicates that our method primarily removes edges that are crucial for the original predictions, resulting in minimal CF explanations.

Author Information

Ana Lucic (Partnership on AI)

Research fellow at the Partnership on AI and PhD student at the University of Amsterdam, working primarily on explainable ML.

Maartje ter Hoeve (University of Amsterdam)
Gabriele Tolomei (University of Rome)
Maarten de Rijke (University of Amsterdam)
Fabrizio Silvestri (Facebook, London, UK)

More from the Same Authors