Research map / Probabilistic and scientific modeling / Variational and generative sampling
AlignFlow: Improving Flow-based Generative Models with Semi-Discrete Optimal Transport
Lingkai Kong, Molei Tao, Yang Liu, Bryan Wang, Jinmiao Fu, Chien-Chih Wang, Huidong Liu
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 minibatch OT coupling with AlignFlow SDOT alignment | flow matching, U-Net · CIFAR-10 (DOPRI5) | FID-50k: 3.82 -> 3.71 | Improved |
| Replace minibatch OT coupling with AlignFlow | flow matching, U-Net · CIFAR-10 (Euler 100 steps) | FID-50k: 4.80 -> 4.72 | Improved |
| Replace minibatch OT coupling with AlignFlow | flow matching, U-Net · CIFAR-10 (Euler 1000 steps) | FID-50k: 3.92 -> 3.79 | Improved |
| Add AlignFlow | Flow Matching, DiT-B/2 · ImageNet256 (NFE=4) | FID-50k: 125.62 -> 93.16 | Improved |
| Add AlignFlow | Consistency Training, DiT-B/2 · ImageNet256 (NFE=1) | FID-50k: 76.37 -> 64.33 | Improved |
| Add AlignFlow | Live Reflow, DiT-B/2 · ImageNet256 (NFE=4) | FID-50k: 94.75 -> 60.23 | Improved |
| Add AlignFlow | Shortcut Models, DiT-B/2 · ImageNet256 (NFE=4) | FID-50k: 33.11 -> 30.31 | Improved |
| Add AlignFlow | Shortcut Models, DiT-B/2 · ImageNet256 (NFE=1) | FID-50k: 46.65 -> 43.92 | Improved |
| Add AlignFlow | MeanFlow SiT-B/4 · ImageNet256 (NFE=1) | FID-50k: 15.53 -> 13.75 | Improved |
| Add AlignFlow | MeanFlow SiT-XL/2 · ImageNet256 (NFE=1) | FID-50k: 3.43 -> 3.23 | Improved |
| Add AlignFlow | MeanFlow SiT-L/2 · ImageNet256 (NFE=1) | FID-50k: 3.84 -> 3.51 | Improved |
| Add rebalance operation to SDOT dual-weight optimization (ablation) | AlignFlow SDOT · CIFAR-10 | FID | No clear effect |
Closest papers by meaning
- Align Your Flow: Scaling Continuous-Time Flow Map Distillation (NeurIPS 2025 · similarity 0.88)
- Faster Inference of Flow-Based Generative Models via Improved Data-Noise Coupling (ICLR 2025 · similarity 0.88)
- Optimal Flow Matching: Learning Straight Trajectories in Just One Step (NeurIPS 2024 · similarity 0.88)
- Improving Flow Matching by Aligning Flow Divergence (ICML 2025 · similarity 0.87)
- Structured Coupling for Flow Matching (NeurIPS 2026 · similarity 0.86)
- On Fitting Flow Models with Large Sinkhorn Couplings (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.
