A Meta-Learning Approach for Training Explainable Graph Neural Networks

TL;DR

This paper introduces MATE, a meta-learning framework that trains GNNs to be more explainable by optimizing for interpretability during training without sacrificing accuracy.

Contribution

The paper presents a novel meta-learning approach that enhances GNN explainability at training time, enabling models to produce more interpretable outputs with any explainer.

Findings

Models trained with MATE are easier to explain across various algorithms.

Explainability improvements do not reduce model accuracy.

The approach is effective on both synthetic and real-world datasets.

Abstract

In this paper, we investigate the degree of explainability of graph neural networks (GNNs). Existing explainers work by finding global/local subgraphs to explain a prediction, but they are applied after a GNN has already been trained. Here, we propose a meta-learning framework for improving the level of explainability of a GNN directly at training time, by steering the optimization procedure towards what we call `interpretable minima'. Our framework (called MATE, MetA-Train to Explain) jointly trains a model to solve the original task, e.g., node classification, and to provide easily processable outputs for downstream algorithms that explain the model's decisions in a human-friendly way. In particular, we meta-train the model's parameters to quickly minimize the error of an instance-level GNNExplainer trained on-the-fly on randomly sampled nodes. The final internal representation relies upon a set of features that can be `better' understood by an explanation algorithm, e.g., another instance of GNNExplainer. Our model-agnostic approach can improve the explanations produced for different GNN architectures and use any instance-based explainer to drive this process. Experiments on synthetic and real-world datasets for node and graph classification show that we can produce models that are consistently easier to explain by different algorithms. Furthermore, this increase in explainability comes at no cost for the accuracy of the model.