Research map / Image generation / Image restoration
Generative Modeling with Optimal Transport Maps
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 OTM rather than the strongest listed baseline | CIFAR10 image generation · CIFAR10 | Inception: 8.87±0.12 -> 7.42±0.06 | Hurt |
| use OTM rather than the strongest listed baseline | CIFAR10 image generation · CIFAR10 | FID: 17.1 -> 21.78 | Hurt |
| use OTM rather than the strongest listed baseline | CelebA image generation · CelebA | FID: 5.2 -> 6.5 | Hurt |
| use OTM rather than the strongest listed baseline | MNIST image generation · MNIST | FID: 3.2 -> 2.4 | Helped |
| use OTM instead of WGAN-GP | CelebA unpaired denoising · CelebA test part C denoising | FID: 25.49 -> 5.92 | Helped |
| use OTM instead of WGAN-GP | CelebA unpaired colorization · CelebA test part C colorization | FID: 7.75 -> 5.65 | Helped |
| use OTM instead of WGAN-GP | CelebA unpaired inpainting · CelebA test part C inpainting | FID: 16.51 -> 8.13 | Helped |
| remove gradient penalty (ablation) | OTM image restoration · CelebA test part C denoising | FID: 10.95 -> 5.92 | Helped |
| remove gradient penalty (ablation) | OTM image restoration · CelebA test part C colorization | FID: 5.66 -> 5.65 | No clear effect |
| remove gradient penalty (ablation) | OTM image restoration · CelebA test part C inpainting | FID: 9.96 -> 8.13 | Helped |
| add gradient optimality regularization with λ=10 instead of using gradient penalty (ablation) | OTM CelebA image generation · CelebA | FID: 7.7 -> 6.50 | Helped |
| vary gradient optimality coefficient λ across tested values (ablation) | OTM CelebA image generation · CelebA | FID: 16.91; 16.22; 16.70; 10.01; 6.50 | Mixed |
Closest papers by meaning
- AE-OT: A NEW GENERATIVE MODEL BASED ON EXTENDED SEMI-DISCRETE OPTIMAL TRANSPORT (ICLR 2020 · similarity 0.90)
- Analyzing and Improving Optimal-Transport-based Adversarial Networks (ICLR 2024 · similarity 0.90)
- Statistically Optimal Generative Modeling with Maximum Deviation from the Empirical Distribution (ICML 2024 · similarity 0.88)
- Overcoming Spurious Solutions in Semi-Dual Neural Optimal Transport: A Smoothing Approach for Learning the Optimal Transport Plan (ICML 2025 · similarity 0.88)
- Improving Neural Optimal Transport via Displacement Interpolation (ICLR 2025 · similarity 0.88)
- Scalable Wasserstein Gradient Flow for Generative Modeling through Unbalanced Optimal Transport (ICML 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.