Skip to yearly menu bar Skip to main content


Poster

GraphChef: Decision-Tree Recipes to Explain Graph Neural Networks

Peter Müller · Lukas Faber · Karolis Martinkus · Roger Wattenhofer

Halle B
[ ]
Fri 10 May 1:45 a.m. PDT — 3:45 a.m. PDT

Abstract:

We propose a new self-explainable Graph Neural Network (GNN) model: GraphChef. GraphChef integrates decision trees into the GNN message passing framework. Given a dataset, GraphChef returns a set of rules (a recipe) that explains each class in the dataset unlike existing GNNs and explanation methods that reason on individual graphs. Thanks to the decision trees, GraphChef recipes are human understandable. We also present a new pruning method to produce small and easy to digest trees. Experiments demonstrate that GraphChef reaches comparable accuracy to not self-explainable GNNs and produced decision trees are indeed small. We further validate the correctness of the discovered recipes on datasets where explanation ground truth is available: Reddit-Binary, MUTAG, BA-2Motifs, BA-Shapes, Tree-Cycle, and Tree-Grid.

Chat is not available.