Research map / Graph learning / Knowledge and temporal graphs
T3former: Temporal Graph Classification with Topological Machine Learning
Md Joshem Uddin, Soham Changani, Baris Coskunuzer
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 attention-based fusion instead of concatenation (ablation) | T3Former · DynHCP-Gender | accuracy: 73.58 -> 75.79 | Improved |
| add attention-based fusion instead of concatenation (ablation) | T3Former · DynHCP-Age | accuracy: 57.78 -> 58.73 | Improved |
| add attention-based fusion instead of concatenation (ablation) | T3Former · DynHCP-Task | accuracy: 90.58 -> 90.76 | Hurt |
| add attention-based fusion instead of concatenation (ablation) | T3Former · Infectious | accuracy: 63.50 -> 68.50 | Improved |
| add attention-based fusion instead of concatenation (ablation) | T3Former · PEMSBAY 3-class | accuracy: 90.47 -> 92.35 | Improved |
| add attention-based fusion instead of concatenation (ablation) | T3Former · PEMS08 3-class | accuracy: 89.92 -> 89.65 | Hurt |
| add topological and spectral transformer branches to static GSAGE (ablation) | T3Former · DynHCP-Age | accuracy: 37.55 -> 58.73 | Improved |
| add topological and spectral transformer branches to static GSAGE (ablation) | T3Former · PEMS04 binary | accuracy: 85.31 -> 96.76 | Improved |
| add static GSAGE branch to Topo-only transformer (ablation) | T3Former · DynHCP-Task | accuracy: 22.10 -> 90.76 | Improved |
| add static GSAGE branch to Topo-only transformer (ablation) | T3Former · MIT | accuracy: 61.90 -> 73.16 | Improved |
| use T3Former instead of GCN+LSTM | T3Former · DynHCP-Task single fold runtime | accuracy: 84.42 -> 5.34 minutes | Improved |
| T3Former vs best temporal baseline | T3Former · Infectious | accuracy: 74.00 vs 68.50 | Hurt |
Closest papers by meaning
- Inductive representation learning on temporal graphs (ICLR 2020 · similarity 0.88)
- S3Former: Sequential, Structural, and Statistical Fusion for Continuous-Time Dynamic Graphs (NeurIPS 2026 · similarity 0.88)
- State Space Models on Temporal Graphs: A First-Principles Study (NeurIPS 2024 · similarity 0.86)
- Temporal Graph Neural Tangent Kernel with Graphon-Guaranteed (NeurIPS 2024 · similarity 0.86)
- Using Time-Aware Graph Neural Networks to Predict Temporal Centralities in Dynamic Graphs (NeurIPS 2024 · similarity 0.85)
- Supra-Laplacian Encoding for Transformer on Dynamic Graphs (NeurIPS 2024 · similarity 0.85)
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.
