Iterative Graph Neural Network Enhancement via Frequent Subgraph Mining of Explanations

TL;DR

This paper introduces EEGL, an iterative method that enhances GNNs for node classification by mining explanations to identify relevant subgraph patterns, leading to improved accuracy and node-distinguishing power.

Contribution

The paper presents a novel explanation-based iterative enhancement method for GNNs, extending the Weisfeiler-Leman algorithm with application-dependent subgraph filtering.

Findings

EEGL outperforms related methods in predictive accuracy

EEGL demonstrates greater node-distinguishing power than vanilla GNNs

Experimental results on synthetic and real data validate effectiveness

Abstract

We formulate an XAI-based model improvement approach for Graph Neural Networks (GNNs) for node classification, called Explanation Enhanced Graph Learning (EEGL). The goal is to improve predictive performance of GNN using explanations. EEGL is an iterative self-improving algorithm, which starts with a learned "vanilla" GNN, and repeatedly uses frequent subgraph mining to find relevant patterns in explanation subgraphs. These patterns are then filtered further to obtain application-dependent features corresponding to the presence of certain subgraphs in the node neighborhoods. Giving an application-dependent algorithm for such a subgraph-based extension of the Weisfeiler-Leman (1-WL) algorithm has previously been posed as an open problem. We present experimental evidence, with synthetic and real-world data, which show that EEGL outperforms related approaches in predictive performance and that it has a node-distinguishing power beyond that of vanilla GNNs. We also analyze EEGL's training dynamics.