Research map / Probabilistic and scientific modeling / Variational and generative sampling
Topological Flow Matching
Kacper Wyrwal, Ismail Ilkan Ceylan, Alexander Tong
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 |
|---|---|---|---|
| replace CFM with TFM (heat-drift reference process) | I-CFM · Earthquakes | 1-Wasserstein distance: 8.37 -> 4.93 | Improved |
| replace CFM with TFM | I-CFM · Traffic flows | 1-Wasserstein distance: 1.72 -> 1.27 | Improved |
| replace CFM with TFM | I-CFM · Brain fMRI | 1-Wasserstein distance: 11.71 -> 6.33 | Improved |
| replace CFM with TFM | OT-CFM · Single-cell | 1-Wasserstein distance: 0.019 -> 0.019 | No clear effect |
| replace CFM with TFM | OT-CFM · Ocean currents | 1-Wasserstein distance: 2.00 -> 1.91 | Improved |
| replace TSBM with simulation-free TFM | TSBM · Traffic flows | 1-Wasserstein distance: 9.92 -> 1.27 | Improved |
| use OT coupling with TFM instead of independent coupling (ablation) | I-TFM · Earthquakes | 1-Wasserstein distance: 4.93 -> 5.53 | Hurt |
| replace CFM with TFM | I-CFM · CIFAR-10 generation | FID: 3.7005 -> 3.6972 | No clear effect |
| replace CFM with TFM | OT-CFM · CIFAR-10 generation | FID: 3.8238 -> 3.8107 | No clear effect |
| use TFM drift instead of topology-aware noising (TAN) (ablation) | I-TAN · Synthetic torus matching | 1-Wasserstein distance: 11.24 -> 8.99 | Improved |
Closest papers by meaning
- Structured Coupling for Flow Matching (NeurIPS 2026 · similarity 0.88)
- Categorical Flow Matching on Statistical Manifolds (NeurIPS 2024 · similarity 0.88)
- Diff2Flow: Training Flow Matching Models via Diffusion Model Alignment (CVPR 2025 · similarity 0.88)
- Tessellations of Semi-Discrete Flow Matching (NeurIPS 2026 · similarity 0.88)
- Metric Flow Matching for Smooth Interpolations on the Data Manifold (NeurIPS 2024 · similarity 0.87)
- Variational Flow Matching for Graph Generation (NeurIPS 2024 · 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.
