Research map / Learning theory and algorithms / Optimal transport
Neural Hamilton--Jacobi Characteristic Flows for Optimal Transport
Yesom Park, Shu Liu, Mo Zhou, Stanley Osher
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 implicit HJ loss with PINN HJ-residual loss (ablation) | NCF · Gaussian OT d=64 | UVP: 1.683 -> 0.858 | Improved |
| replace implicit HJ loss with PINN HJ-residual loss (ablation) | NCF · Gaussian OT d=2 | UVP: 0.080 -> 0.010 | Improved |
| use NCF instead of MM-v1 (dual min-max ICNN) | NCF · Gaussian OT d=32 | UVP: 0.374 vs 0.436 | Hurt |
| use NCF instead of MM-v1 | NCF · Gaussian OT d=2-16 | UVP: 0.161 -> 0.010 (d=2) | Improved |
| use NCF instead of NOT | NCF · Gaussian OT d=64 | UVP: 196.831 -> 0.858 | Improved |
| use NCF instead of MM-v1 for color transfer | NCF · Winter-Summer color transfer | histogram intersection: 0.7295 -> 0.8914 | Improved |
| use NCF instead of NOT for color transfer | NCF · Gogh-Photograph color transfer | EMD: 0.0008 -> 0.0003 | Improved |
| use NCF instead of GNOT for class-conditional OT | NCF · Fashion MNIST to MNIST | accuracy (%): 83.22 -> 83.42 | No clear effect |
| use NCF instead of GNOT for class-conditional OT | NCF · Fashion MNIST to MNIST | FID: 5.26 -> 18.27 | Hurt |
Closest papers by meaning
- Computing Optimal Transport Maps and Wasserstein Barycenters Using Conditional Normalizing Flows (ICML 2025 · similarity 0.88)
- Progressive Entropic Optimal Transport Solvers (NeurIPS 2024 · similarity 0.86)
- A Computational Framework for Solving Wasserstein Lagrangian Flows (ICML 2024 · similarity 0.86)
- Improving Neural Optimal Transport via Displacement Interpolation (ICLR 2025 · similarity 0.86)
- Flow Matching from Viewpoint of Proximal Operators (NeurIPS 2026 · similarity 0.85)
- Expectile Regularization for Fast and Accurate Training of Neural Optimal Transport (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.
