Counterfactual Explanations for Hypergraph Neural Networks

TL;DR

The paper tackles the interpretability of hypergraph neural networks by introducing CF-HyperGNNExplainer, a counterfactual explanation method that generates minimal, actionable edits to hypergraphs to flip predictions. It presents two perturbation variants, NHP and HP, to operate at different granularities and demonstrates their effectiveness on three citation-network datasets, achieving high sparsity and competitive accuracy. By adapting counterfactual explanations to the hypergraph setting, the work shows that higher-order interactions are pivotal in HGNN decisions and provides a principled, efficient framework compared to graph-based baselines. This advance enables more trustworthy deployment of HGNNs in high-stakes domains and lays groundwork for future extensions to richer perturbations and hypergraph-level tasks.

Abstract

Hypergraph neural networks (HGNNs) effectively model higher-order interactions in many real-world systems but remain difficult to interpret, limiting their deployment in high-stakes settings. We introduce CF-HyperGNNExplainer, a counterfactual explanation method for HGNNs that identifies the minimal structural changes required to alter a model's prediction. The method generates counterfactual hypergraphs using actionable edits limited to removing node-hyperedge incidences or deleting hyperedges, producing concise and structurally meaningful explanations. Experiments on three benchmark datasets show that CF-HyperGNNExplainer generates valid and concise counterfactuals, highlighting the higher-order relations most critical to HGNN decisions.

Figures (1)

• Figure 1: Temporal evolution of publications containing the exact phrase "hypergraph neural network".