AutoHGNN: Robust and efficient neural architecture search for hypergraph neural networks

2026-09-27

Knowledge-Based Systems

AI Summary

This study addresses the limitations of manually designed hypergraph neural networks, which are time-consuming to construct and often fail to adequately capture higher-order relationships. To this end, this work proposes AutoHGNN, the first neural architecture search framework tailored for hypergraphs. Methodologically, it introduces a Hyper Interaction Module (HIM) specifically designed to accommodate the intrinsic properties of hypergraph data, along with a Hypergraph Structure-Topology-preserving Differentiable (HyperSTD) criterion to optimize the differentiable search process. This work achieves automated discovery of hypergraph network architectures. Extensive experiments on multiple benchmark datasets demonstrate that AutoHGNN significantly outperforms existing handcrafted and automatically searched baseline models in both classification accuracy and search efficiency.

Abstract

Hypergraph neural networks have achieved significant success in recent years. However, manual architecture crafting is labor-intensive and often fails to capture complex, higher-order relations, making the automation of hypergraph neural network structure design crucial. To improve the automation and adaptability of hypergraph learning, this paper proposes AutoHGNN, a neural architecture search framework tailored for hypergraph neural networks. First, we introduce a Hyper-Interaction Module (HIM) into the search space to address the mismatch between conventional graph neural network designs and hypergraph data. Second, we propose Hypergraph Stable Topological Distance (HyperSTD) as a structural selection criterion to identify architectures that best preserve the intrinsic structural affinities of the original hypergraph during differentiable search. Extensive experiments on various benchmark datasets demonstrate that AutoHGNN consistently outperforms manually designed and automatically searched baselines in classification accuracy and time efficiency, proving that the discovered architectures are significantly more effective.

Problem

• Hypergraph Neural Networks
• Neural Architecture Search
• Higher-order Relations
• Automation

Innovation

• Neural Architecture Search
• Hypergraph Neural Networks
• Hyper-Interaction Module
• Hypergraph Stable Topological Distance
• Differentiable Search