GraphSVX: Shapley Value Explanations for Graph Neural Networks

TL;DR

GraphSVX introduces a novel, model-agnostic explanation method for GNNs that uses Shapley Values to fairly attribute contributions of features and nodes, improving interpretability.

Contribution

The paper presents GraphSVX, a new explanation technique for GNNs that extends Shapley Values to graph data, offering state-of-the-art performance and theoretical properties.

Findings

GraphSVX outperforms baseline explanation methods.

It provides fair and human-centric explanations.

The method is applicable to real-world and synthetic datasets.

Abstract

Graph Neural Networks (GNNs) achieve significant performance for various learning tasks on geometric data due to the incorporation of graph structure into the learning of node representations, which renders their comprehension challenging. In this paper, we first propose a unified framework satisfied by most existing GNN explainers. Then, we introduce GraphSVX, a post hoc local model-agnostic explanation method specifically designed for GNNs. GraphSVX is a decomposition technique that captures the "fair" contribution of each feature and node towards the explained prediction by constructing a surrogate model on a perturbed dataset. It extends to graphs and ultimately provides as explanation the Shapley Values from game theory. Experiments on real-world and synthetic datasets demonstrate that GraphSVX achieves state-of-the-art performance compared to baseline models while presenting core theoretical and human-centric properties.