SEEN: Sharpening Explanations for Graph Neural Networks using Explanations from Neighborhoods

TL;DR

SEEN enhances the explanation quality of graph neural network predictions by aggregating auxiliary explanations from neighboring nodes, improving accuracy without modifying the original graph or explainability methods.

Contribution

Introduces SEEN, a post hoc method that improves GNN explanation quality by aggregating neighboring node explanations, compatible with various explainability techniques.

Findings

Explanation accuracy improved by up to 12.71%.

Correlation observed between auxiliary explanations and explanation quality.

Method applicable with diverse explainability techniques.

Abstract

Explaining the foundations for predictions obtained from graph neural networks (GNNs) is critical for credible use of GNN models for real-world problems. Owing to the rapid growth of GNN applications, recent progress in explaining predictions from GNNs, such as sensitivity analysis, perturbation methods, and attribution methods, showed great opportunities and possibilities for explaining GNN predictions. In this study, we propose a method to improve the explanation quality of node classification tasks that can be applied in a post hoc manner through aggregation of auxiliary explanations from important neighboring nodes, named SEEN. Applying SEEN does not require modification of a graph and can be used with diverse explainability techniques due to its independent mechanism. Experiments on matching motif-participating nodes from a given graph show great improvement in explanation accuracy of up to 12.71% and demonstrate the correlation between the auxiliary explanations and the enhanced explanation accuracy through leveraging their contributions. SEEN provides a simple but effective method to enhance the explanation quality of GNN model outputs, and this method is applicable in combination with most explainability techniques.