arXiv：cs.LG· Shihan Feng, Xin Zheng, Shiyi Yang, Ren Wang, Chudi Zhong, Can Chen·· 4 小时前AI 评分30
超图神经加性网络 HGNAN：用神经加性模型实现可解释超图学习
Interpretable Hypergraph Learning via Neural Additive Models
AI 导读
研究者提出超图神经加性网络（HGNAN），将经典神经加性模型扩展到高阶关系数据，通过特征级非线性分解与超图感知结构聚合，为节点级和超边级任务提供透明预测。在基准数据集上，HGNAN 性能与当前最先进的超图学习方法相当，同时具备内在可解释性。该工作以 arXiv:2610.07458 发布，共 16 页、8 图、13 表。
正文
Abstract:Hypergraphs offer a natural framework for modeling networked data, where dependencies among entities are governed by higher-order interactions. While hypergraph learning methods such as hypergraph neural networks have demonstrated remarkable predictive performance, most existing approaches rely on black-box message-passing architectures, making it difficult to disentangle the contributions of node attributes and higher-order structural information. To address this challenge, we introduce the hypergraph neural additive network (HGNAN), an inherently interpretable framework for learning on hypergraph-structured data. HGNAN extends classical neural additive models to higher-order relational data by integrating feature-wise nonlinear decomposition with hypergraph-aware structural aggregation, enabling transparent prediction for both node- and hyperedge-level tasks. Extensive experiments on benchmark datasets demonstrate that HGNAN achieves performance comparable with state-of-the-art hypergraph learning methods while providing intrinsic and meaningful interpretability.
| Comments: | 16 pages, 8 figures, 13 tables |
| Subjects: | Machine Learning (cs.LG); Social and Information Networks (cs.SI) |
| Cite as: | arXiv:2610.07458 [cs.LG] |
| (or arXiv:2610.07458v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.07458 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Can Chen [view email]
[v1]
Mon, 5 Oct 2026 22:07:40 UTC (1,492 KB)
来源：arXiv：cs.LG · arxiv.org