Research map / Graph learning / Graph self-supervised learning
Beyond Convolution: Advancing Hypergraph Neural Networks with Hypergraph U-Nets
Fuli Wang, Wei Qian, Daniel L. Lau, Gonzalo R. Arce
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 PHPool to HGAT backbone (ablation) | HGAT · MUTAG | accuracy: 74.53 -> 86.73 | Improved |
| add PHPool to HGNN backbone (ablation) | HGNN · IMDB-BINARY | accuracy: 72.78 -> 74.88 | Improved |
| add PHPool to HGXConv backbone (ablation) | HGXConv · IMDB-MULTI | accuracy: 50.76 -> 55.31 | Improved |
| replace SEPool with PHPool on HGXConv (ablation) | SEPool+HGXConv · COLLAB | accuracy: 80.73 -> 83.26 | Improved |
| replace TopKPool with PHPool on HGXConv (ablation) | TopKPool+HGXConv · COLLAB | accuracy: 68.55 -> 83.26 | Improved |
| replace attention backbone HGAT with HGXConv under PHPool (ablation) | PHPool+HGAT · MUTAG | accuracy: 86.73 -> 86.42 | Hurt |
| use HGXConv instead of HGNN under PHPool (ablation) | PHPool+HGNN · D&D | accuracy: 77.93 -> 79.88 | Improved |
| use HGXConv instead of GCN-style convolution alone (no pooling) | HGXConv · MUTAG | accuracy: 73.37 vs 74.53 (HGAT) | Mixed |
| increase pooling depth beyond 3 (ablation) | HGUN encoder · COLLAB | accuracy: sharp degradation at depth 5 | Hurt |
| use TopK node-drop pooling at depth 4 (ablation) | TopK · D&D | accuracy: dropping 11 points | Hurt |
| use HGUN instead of BWGNN | BWGNN · Yelp anomaly detection | F1-macro: 0.7232 -> 0.7732 | Improved |
| use HGUN instead of HGXConv without U-Net (ablation) | HGXConv · Yelp anomaly detection | F1-macro: 0.7227 -> 0.7732 | Improved |
Closest papers by meaning
- Graph U-Nets (ICML 2019 · similarity 0.89)
- SSHPool: The Separated Subgraph-based Hierarchical Pooling (AAAI 2026 · similarity 0.88)
- Hypergraph Learning for Unsupervised Graph Alignment via Optimal Transport (AAAI 2025 · similarity 0.87)
- When Hypergraph Meets Heterophily: New Benchmark Datasets and Baseline (AAAI 2025 · similarity 0.87)
- K-hop Hypergraph Neural Network: A Comprehensive Aggregation Approach (AAAI 2025 · similarity 0.87)
- Deep Hypergraph Neural Networks with Tight Framelets (AAAI 2025 · 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.
