Research map / Graph learning / Graph self-supervised learning
Brain Network Transformer
Xuan Kan, Wei Dai, Hejie Cui, Zilong Zhang, Ying Guo, Carl Yang
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 OCRead readout in place of CONCAT (ablation) | VanillaTF · ABIDE | AUROC: 76.4 -> 80.2 | Improved |
| add OCRead readout in place of CONCAT (ablation) | VanillaTF · ABCD | AUROC: 94.3 -> 96.2 | Improved |
| add OCRead readout in place of CONCAT (ablation) | SAN · ABIDE | AUROC: 71.3 -> 70.6 | Hurt |
| add OCRead readout in place of CONCAT (ablation) | Graphormer · ABIDE | AUROC: 63.5 -> 64.9 | Improved |
| add OCRead readout in place of CONCAT (ablation) | SAN · ABCD | AUROC: 90.1 -> 91.2 | Improved |
| use OCRead instead of MAX readout (ablation) | VanillaTF · ABCD | AUROC: 94.4 -> 96.2 | Improved |
| add identity feature to connection profile node features (ablation) | VanillaTF · ABIDE | AUROC: 76.4 -> 75.4 | No clear effect |
| add eigen feature to connection profile node features (ablation) | VanillaTF · ABCD | AUROC: 94.3 -> 94.0 | No clear effect |
| use BrainNetTF (connection profile + vanilla attention + OCRead) instead of SAN (ablation) | graph Transformer · ABIDE | AUROC: 71.3 -> 80.2 | Improved |
| use BrainNetTF instead of FBNETGNN learnable-graph model (ablation) | brain network classifier · ABCD | AUROC: 94.5 -> 96.2 | Improved |
| replace random cluster initialization with orthonormal initialization in OCRead (ablation) | BrainNetTF · ABIDE and ABCD | AUROC | Mixed |
| increase number of clusters K from 10 to 100 in OCRead (ablation) | BrainNetTF · ABIDE and ABCD | AUROC | Hurt |
Closest papers by meaning
- Long-range Brain Graph Transformer (NeurIPS 2024 · similarity 0.90)
- Biologically Plausible Brain Graph Transformer (ICLR 2025 · similarity 0.89)
- Do We Really Need Message Passing in Brain Network Modeling? (ICML 2025 · similarity 0.87)
- BrainOOD: Out-of-distribution Generalizable Brain Network Analysis (ICLR 2025 · similarity 0.86)
- Functional Connectomes of Neural Networks (AAAI 2025 · similarity 0.86)
- Information Bottleneck-Guided Adaptive Hypergraph Transformer for Brain Disease Diagnosis (NeurIPS 2026 · similarity 0.86)
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.
