Towards Self-Explainable Graph Neural Network

TL;DR

This paper introduces a novel framework for self-explainable Graph Neural Networks that provides both predictions and explanations by identifying similar labeled nodes, addressing the limitations of post-hoc explanation methods.

Contribution

The paper proposes a new self-explainable GNN framework that simultaneously predicts and explains node classifications using an interpretable similarity module.

Findings

Effective in real-world datasets

Accurate explanations via nearest labeled nodes

Improves transparency of GNN predictions

Abstract

Graph Neural Networks (GNNs), which generalize the deep neural networks to graph-structured data, have achieved great success in modeling graphs. However, as an extension of deep learning for graphs, GNNs lack explainability, which largely limits their adoption in scenarios that demand the transparency of models. Though many efforts are taken to improve the explainability of deep learning, they mainly focus on i.i.d data, which cannot be directly applied to explain the predictions of GNNs because GNNs utilize both node features and graph topology to make predictions. There are only very few work on the explainability of GNNs and they focus on post-hoc explanations. Since post-hoc explanations are not directly obtained from the GNNs, they can be biased and misrepresent the true explanations. Therefore, in this paper, we study a novel problem of self-explainable GNNs which can simultaneously give predictions and explanations. We propose a new framework which can find $K$-nearest labeled nodes for each unlabeled node to give explainable node classification, where nearest labeled nodes are found by interpretable similarity module in terms of both node similarity and local structure similarity. Extensive experiments on real-world and synthetic datasets demonstrate the effectiveness of the proposed framework for explainable node classification.