View-based Explanations for Graph Neural Networks

TL;DR

This work introduces GVEX, a graph-view based approach to explain GNNs for graph classification by generating label-specific, queryable explanations that combine lower-tier subgraph explanations with higher-tier graph patterns. It formalizes the Explanation View Generation problem, proves hardness results, and leverages submodular optimization to develop two algorithms: ApproxGVEX with a 1/2-approximation and StreamGVEX with a 1/4-approximation in streaming, both providing explainability and coverage guarantees. The methods are model-agnostic, scalable, and capable of producing concise, pattern-based summaries that support domain-oriented queries, demonstrated through extensive experiments on real-world biomedical and social datasets. The results show improved faithfulness, conciseness, and efficiency compared with existing GNN explainers, highlighting practical impact for explainable graph analytics and drug discovery workflows.

Abstract

Generating explanations for graph neural networks (GNNs) has been studied to understand their behavior in analytical tasks such as graph classification. Existing approaches aim to understand the overall results of GNNs rather than providing explanations for specific class labels of interest, and may return explanation structures that are hard to access, nor directly queryable.We propose GVEX, a novel paradigm that generates Graph Views for EXplanation. (1) We design a two-tier explanation structure called explanation views. An explanation view consists of a set of graph patterns and a set of induced explanation subgraphs. Given a database G of multiple graphs and a specific class label l assigned by a GNN-based classifier M, it concisely describes the fraction of G that best explains why l is assigned by M. (2) We propose quality measures and formulate an optimization problem to compute optimal explanation views for GNN explanation. We show that the problem is $Σ^2_P$-hard. (3) We present two algorithms. The first one follows an explain-and-summarize strategy that first generates high-quality explanation subgraphs which best explain GNNs in terms of feature influence maximization, and then performs a summarization step to generate patterns. We show that this strategy provides an approximation ratio of 1/2. Our second algorithm performs a single-pass to an input node stream in batches to incrementally maintain explanation views, having an anytime quality guarantee of 1/4 approximation. Using real-world benchmark data, we experimentally demonstrate the effectiveness, efficiency, and scalability of GVEX. Through case studies, we showcase the practical applications of GVEX.

Figures (14)

• Figure 1: $$GNN-based drug classification, with patterns and induced subgraphs that help understand the results.
• Figure 2: An explanation view for a single class label: explanation subgraphs and patterns
• Figure 3: "Explain-and-Summarize": an illustration
• Figure 4: Incremental generation of explanation views
• Figure 5: The Fidelity+ comparison across various GNN explainers under different configuration constraints

Theorems & Definitions (6)

• lemma 1
• theorem 1
• lemma 2
• theorem 2
• lemma 3
• theorem 3