Research map / Probabilistic and scientific modeling / Variational and generative sampling
WFR-FM: Simulation-Free Dynamic Unbalanced Optimal Transport
Qiangwei Peng, Zihan Wang, Junda Ying, Yuhao Sun, Qing Nie, Lei Zhang, Tiejun Li, Peijie Zhou
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 |
|---|---|---|---|
| use WFR-FM instead of VGFM | VGFM · Synthetic Dyngen (5D) | W1 distance: 0.598 -> 0.135 | Improved |
| use WFR-FM instead of DeepRUOT | DeepRUOT · Synthetic Gene (2D) | W1 distance: 0.043 -> 0.019 | Improved |
| use WFR-FM instead of TIGON on high-dimensional data | TIGON · Synthetic Gaussian (1000D) | W1 distance: 2.263 -> 2.233 | Improved |
| use WFR-FM instead of UOT-FM | UOT-FM · Synthetic Gaussian (1000D) | relative mass error: 0.033 -> 0.044 | Hurt |
| use WFR-FM instead of DeepRUOT | DeepRUOT · Hold-one-out EB (50D) | mean W1 over held-out times: 10.075 -> 10.157 | Hurt |
| use WFR-FM instead of VGFM | VGFM · Hold-one-out EMT (10D) | mean W1 over held-out times: 0.301 -> 0.298 | Improved |
| use WFR-FM instead of DeepRUOT | DeepRUOT · Hold-one-out Mouse (50D) | mean W1 over held-out times: 6.847 -> 6.586 | Improved |
| use WFR-FM instead of VGFM | VGFM · Hold-one-out CITE (50D) | mean W1 over held-out times: 37.386 -> 37.221 | Improved |
| use WFR-FM instead of Var-RUOT for path action | Var-RUOT · Path action, Gene | path action: 1.208 -> 1.305 | Hurt |
| use WFR-FM instead of DeepRUOT for path action | DeepRUOT · Path action, Dyngen | path action: 14.070 -> 9.410 | Improved |
| use WFR-FM instead of Action Matching for growth | Action Matching · Simulation Gene growth rate | growth rate correlation: 0.5851 -> 0.9913 | Improved |
| use WFR-FM instead of TIGON for growth | TIGON · Simulation Gene growth rate | growth rate correlation: 0.9705 -> 0.9913 | Improved |
Later papers that compared with this method
2 later accepted papers report WFR-FM as a baseline in their results tables; 0 of them report a better number on most of the results they share with it.
- Beyond Continuity: Simulation-free Reconstruction of Discrete Branching Dynamics from Single-cell Snapshots (ICML 2026 · better on 4 of 10 shared results)
- WFR-MFM: One-Step Inference for Dynamic Unbalanced Optimal Transport (ICML 2026 · better on 1 of 6 shared results)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Joint Velocity-Growth Flow Matching for Single-Cell Dynamics Modeling (NeurIPS 2025 · similarity 0.89)
- TracingFlow: A Simulation-Free Trajectory Inference Framework Based on Second-Order Dynamics (NeurIPS 2026 · similarity 0.88)
- Generalized Wasserstein Flow Matching: Transport Plans, Everywhere, All at Once (NeurIPS 2026 · similarity 0.87)
- Curly Flow Matching for Learning Non-gradient Field Dynamics (NeurIPS 2025 · similarity 0.87)
- Flow Matching from Viewpoint of Proximal Operators (NeurIPS 2026 · similarity 0.87)
- Metric Flow Matching for Smooth Interpolations on the Data Manifold (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.
