Research map / Graph learning / GNN architectures
Topological Relational Learning on Graphs
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Remove TIMR (topology-induced multiedge representation) (ablation) | TRI-GNN_D · ACTIVSg200 | accuracy: 82.93 -> 86.18 | Helped |
| Remove STAN (subgraph-topology-neighborhood feature aggregation) (ablation) | TRI-GNN_D · ACTIVSg200 | accuracy: 83.00 -> 86.18 | Helped |
| Use TIMR and STN only (drop neighborhood features) (ablation) | TRI-GNN_D · ACTIVSg500 | accuracy: 93.69 -> 98.11 | Helped |
| Remove TIMR (ablation) | TRI-GNN_D · Cora-ML | accuracy: 84.26 -> 84.98 | Helped |
| Remove TIMR (ablation) | TRI-GNN_D · CiteSeer | accuracy: 73.11 -> 73.32 | No clear effect |
| Remove TIMR (ablation) | TRI-GNN_A · PubMed | accuracy: 79.77 -> 79.80 | No clear effect |
| Remove STAN (ablation) | TRI-GNN_A · CiteSeer | accuracy: 73.10 -> 73.45 | No clear effect |
| Use node degree instead of node attributes as filtration function | TRI-GNN · ACTIVSg200 | accuracy: 84.93 -> 86.18 | Helped |
| Use node attributes instead of node degree as filtration function | TRI-GNN · IEEE 118-Bus | accuracy: 82.67 -> 82.80 | Helped |
| Reduce training set from 100% to 80% (ablation) | TRI-GNN_D · Cora-ML | accuracy: 84.98 -> 84.00 | Hurt |
| Reduce training set from 100% to 50% (ablation) | TRI-GNN_D · Cora-ML | accuracy: 84.98 -> 73.20 | Hurt |
| Increase layers from 2 to 32 (ablation) | TRI-GNN_D · Cora-ML | accuracy: 84.98 -> 72.87 | Hurt |
Closest papers by meaning
- Weisfeiler and Lehman Go Paths: Learning Topological Features via Path Complexes (AAAI 2024 · similarity 0.91)
- Topological Blindspots: Understanding and Extending Topological Deep Learning Through the Lens of Expressivity (ICLR 2025 · similarity 0.91)
- TopoTune: A Framework for Generalized Combinatorial Complex Neural Networks (ICML 2025 · similarity 0.90)
- TopInG: Topologically Interpretable Graph Learning via Persistent Rationale Filtration (ICML 2025 · similarity 0.90)
- A Topological Perspective on Demystifying GNN-Based Link Prediction Performance (ICLR 2024 · similarity 0.89)
- TopoGCL: Topological Graph Contrastive Learning (AAAI 2024 · similarity 0.89)
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.