HyGEN: Regularizing Negative Hyperedge Generation for Accurate Hyperedge Prediction
Song Kyung Yu, Da Eun Lee, Yunyong Ko, Sang-Wook Kim
TL;DR
HyGEN tackles hyperedge prediction under severe data sparsity by addressing two neglected issues in negative sampling: insufficient guidance for negative generation and the risk of false negatives. It introduces a positive-guided negative hyperedge generator and a cosine-similarity–based regularization term, trained adversarially with a discriminator that uses max-min pooling to score hyperedge candidates. Across six real-world hypergraphs, HyGEN consistently surpasses four state-of-the-art baselines in AUROC and AP, with ablation studies confirming the value of both core components and a notable robustness to regularization hyperparameters. The work provides code and datasets to support reproducibility and practical adoption in hypergraph learning tasks.
Abstract
Hyperedge prediction is a fundamental task to predict future high-order relations based on the observed network structure. Existing hyperedge prediction methods, however, suffer from the data sparsity problem. To alleviate this problem, negative sampling methods can be used, which leverage non-existing hyperedges as contrastive information for model training. However, the following important challenges have been rarely studied: (C1) lack of guidance for generating negatives and (C2) possibility of producing false negatives. To address them, we propose a novel hyperedge prediction method, HyGEN, that employs (1) a negative hyperedge generator that employs positive hyperedges as a guidance to generate more realistic ones and (2) a regularization term that prevents the generated hyperedges from being false negatives. Extensive experiments on six real-world hypergraphs reveal that HyGEN consistently outperforms four state-of-the-art hyperedge prediction methods.