Research map / Graph learning / Graph self-supervised learning
Self-Supervised Hypergraph Learning with Substructure Awareness for Hyperedge Prediction
Ming Li, Huiting Wang, Yuting Chen, Lu Bai, Lixin Cui, Feilong Cao, Ke Lv
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add SCHA and A3CL to baseline (ablation) | S3Hyper · Cora | average AUROC: 0.697 -> 0.829 | Improved |
| add SCHA (substructure-contextualized aggregator) (ablation) | S3Hyper w/o SCHA · Cora | average AUROC: 0.724 -> 0.829 | Improved |
| add A3CL (tri-directional contrastive learning) (ablation) | S3Hyper w/o A3CL · Citeseer | average AUROC: 0.833 -> 0.861 | Improved |
| add A3CL (ablation) | S3Hyper w/o A3CL · NDC-class | average AUROC: 0.857 -> 0.862 | Improved |
| add SCHA (ablation) | S3Hyper w/o SCHA · NDC-class | average AUROC: 0.778 -> 0.862 | Improved |
| add A3CL (ablation) | S3Hyper w/o A3CL · DBLP SNS | AUROC: 0.849 -> 0.880 | No clear effect |
| replace mean/max-min/attention aggregation with SCHA (ablation) | S3Hyper · Citeseer | AUROC: 0.85 -> 0.86 (attention vs SCHA) | Improved |
| replace best other aggregator with SCHA (ablation) | S3Hyper · NDC-class / DBLP | AUROC: SCHA 0.86 / 0.82 | Improved |
| vary contrastive loss weights lambda in [0.1, 1.0] (ablation) | S3Hyper · NDC-class, Cora | AUROC | No clear effect |
| S3Hyper vs CASH on Citeseer | S3Hyper · Citeseer SNS | AUROC: CASH 0.925 vs S3Hyper 0.940 (NHP 0.991 best) | Hurt |
| use S3Hyper instead of CASH | S3Hyper · NDC-class | average AUROC: 0.752 -> 0.862 | Improved |
Closest papers by meaning
- HyperNoRA: Hyperedge Prediction via Node-Level Relation-Aware Self-Supervised Hypergraph Learning (AAAI 2026 · similarity 0.93)
- Hypergraph Learning for Unsupervised Graph Alignment via Optimal Transport (AAAI 2025 · similarity 0.88)
- Hypergraph Joint Representation Learning for Hypervertices and Hyperedges via Cross Expansion (AAAI 2024 · similarity 0.88)
- SPHINX: Structural Prediction using Hypergraph Inference Network (ICML 2025 · similarity 0.87)
- Disentangling Hyperedges through the Lens of Category Theory (NeurIPS 2025 · similarity 0.87)
- HypeBoy: Generative Self-Supervised Representation Learning on Hypergraphs (ICLR 2024 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
