Research map / Graph learning / GNN architectures
Torsion Graph Neural Networks
Cong Shen, Xiang Liu, Jiawei Luo, Kelin Xia
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 |
|---|---|---|---|
| weight message passing with analytic torsion (TorGNN vs GCN) | GCN · link prediction average of 16 networks | AUC: 0.9144 -> 0.9372 | Improved |
| weight message passing with analytic torsion (TorGNN vs GCN) | GCN · link prediction average of 16 networks | AUPR: 0.8649 -> 0.9319 | Improved |
| analytic torsion weight versus best baseline | KGIN · link prediction average of 16 networks | AUC: 0.9011 -> 0.9372 | Improved |
| use 2-hop neighborhood for local complex (l_sub=2) (ablation) | TorGNN(1,1) · Drugbank_DTI link prediction | AUC: 0.9380 -> 0.9651 | Improved |
| use 2-hop neighborhood for local complex (l_sub=2) (ablation) | TorGNN(1,1) · ChG-Miner link prediction | AUC: 0.9533 -> 0.9583 | Improved |
| use 2-hop neighborhood for local complex (l_sub=2) (ablation) | TorGNN(1,1) · facebook_comp link prediction | AUC: 0.8974 -> 0.8573 | Hurt |
| use 2-hop neighborhood for local complex (l_sub=2) (ablation) | TorGNN(1,1) · p2p-Gnutella06 link prediction | AUC: 0.8464 -> 0.9051 | Improved |
| add triangles (n=2) to local complex (ablation) | TorGNN(1,1) · Drugbank_DTI link prediction | AUC: 0.9380 -> 0.9306 | Hurt |
| add triangles (n=2) to local complex (ablation) | TorGNN(1,1) · CA-GrQc link prediction | AUC: 0.9573 -> 0.9328 | Hurt |
| add triangles (n=2) to local complex (ablation) | TorGNN(1,1) · Pubmed node classification | accuracy: 0.865 -> 0.872 | Improved |
| add triangles (n=2) to local complex (ablation) | TorGNN(1,1) · Cora node classification | accuracy: 0.832 -> 0.823 | Hurt |
| use 2-hop neighborhood for local complex (l_sub=2) (ablation) | TorGNN(1,1) · Citeseer node classification | accuracy: 0.748 -> 0.749 | No clear effect |
Closest papers by meaning
- Are High-Degree Representations Really Unnecessary in Equivariant Graph Neural Networks? (NeurIPS 2024 · similarity 0.86)
- On the Completeness of Invariant Geometric Deep Learning Models (ICLR 2025 · similarity 0.85)
- A Topological Perspective on Demystifying GNN-Based Link Prediction Performance (ICLR 2024 · similarity 0.85)
- A Manifold Perspective on the Statistical Generalization of Graph Neural Networks (ICML 2025 · similarity 0.85)
- What do Graph Neural Networks learn? Insights from Tropical Geometry (NeurIPS 2024 · similarity 0.85)
- What Is Missing For Graph Homophily? Disentangling Graph Homophily For Graph Neural Networks (NeurIPS 2024 · similarity 0.84)
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.
