Research map / Probabilistic and scientific modeling / Variational and generative sampling
COT-FM: Cluster-wise Optimal Transport Flow Matching
Chiensheng Chiang, Kuan-Hsun Tu, Jia-Wei Liao, Cheng-Fu Chou, Tsung-Wei Ke
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 cluster-wise source distributions with random coupling (ablation) | Rectified Flow · CIFAR-10 1-step | FID: 378.0 -> 296.0 | Improved |
| add batch optimal coupling within each cluster (ablation) | Rectified Flow + clustering · CIFAR-10 1-step | FID: 296.0 -> 205.0 | Improved |
| apply COT-FM to Rectified Flow | Rectified Flow · CIFAR-10 10-step | FID: 12.6 -> 8.23 | Improved |
| apply COT-FM to Rectified Flow | Rectified Flow · CIFAR-10 50-step | FID: 4.45 -> 3.97 | Improved |
| apply COT-FM to MeanFlow | MeanFlow · CIFAR-10 1-step | FID: 2.92 -> 2.60 | Improved |
| apply COT-FM to 2-Rectified Flow | 2-Rectified Flow · CIFAR-10 1-step | FID: 12.21 -> 10.91 | Improved |
| increase alternating iterations from 2 to 3 (ablation) | COT-FM · CIFAR-10 50-step | FID: 3.97 -> 4.17 | Hurt |
| sample clusters proportional to size instead of uniformly (ablation) | COT-FM · CIFAR-10 50-step | FID: 4.26 -> 3.97 | Improved |
| increase number of clusters K from 5 to 100 (ablation) | COT-FM · CIFAR-10 50-step | FID: 4.56 -> 3.97 | Improved |
| increase number of clusters K from 100 to 120 (ablation) | COT-FM · CIFAR-10 50-step | FID: 3.97 -> 4.06 | Hurt |
| apply COT-FM to Rectified Flow with SiT-B/2 | SiT-B/2 · ImageNet 256 1 NFE | FID: 264.36 -> 231.99 | Improved |
| train Rectified Flow longer for equal compute to reverse ODE (ablation) | Rectified Flow · CIFAR-10 50-step | FID: 4.45 -> 4.49 | No clear effect |
Closest papers by meaning
- Optimal Flow Matching: Learning Straight Trajectories in Just One Step (NeurIPS 2024 · similarity 0.90)
- Reflected Flow Matching (ICML 2024 · similarity 0.87)
- Faster Inference of Flow-Based Generative Models via Improved Data-Noise Coupling (ICLR 2025 · similarity 0.87)
- Flow Matching from Viewpoint of Proximal Operators (NeurIPS 2026 · similarity 0.87)
- Structured Coupling for Flow Matching (NeurIPS 2026 · similarity 0.87)
- Beckmann Transport Models: From Autonomous Flows to One-Step Maps (NeurIPS 2026 · 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.
