Explaining Hypergraph Neural Networks: From Local Explanations to Global Concepts

TL;DR

This paper introduces SHypX, a novel model-agnostic explainer for hypergraph neural networks that provides both local and global explanations, improving interpretability and fidelity in hypergraph models.

Contribution

It presents the first post-hoc explanation method for hypergraph neural networks, enabling instance-level and model-level interpretability with high fidelity.

Findings

SHypX achieves 25% higher fidelity than baselines.

It effectively balances faithfulness and conciseness.

Demonstrated on eight diverse hypergraph datasets.

Abstract

Hypergraph neural networks are a class of powerful models that leverage the message passing paradigm to learn over hypergraphs, a generalization of graphs well-suited to describing relational data with higher-order interactions. However, such models are not naturally interpretable, and their explainability has received very limited attention. We introduce SHypX, the first model-agnostic post-hoc explainer for hypergraph neural networks that provides both local and global explanations. At the instance-level, it performs input attribution by discretely sampling explanation subhypergraphs optimized to be faithful and concise. At the model-level, it produces global explanation subhypergraphs using unsupervised concept extraction. Extensive experiments across four real-world and four novel, synthetic hypergraph datasets demonstrate that our method finds high-quality explanations which can target a user-specified balance between faithfulness and concision, improving over baselines by 25 percent points in fidelity on average.