Research map / Learning theory and algorithms / Optimal transport
A Gradual, Semi-Discrete Approach to Generative Network Training via Explicit Wasserstein Minimization
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 alternating OTS and FIT instead of fitting once to the initial transport plan (ablation) | Proposed generative method · MNIST | visual quality | Helped |
| Use alternating OTS and FIT instead of fitting once to the initial transport plan (ablation) | Proposed generative method · synthetic example | manifold smoothness | Helped |
| Apply the proposed method rather than WAE | MLP generator · MNIST training | WD: 90.91 -> 66.68 | Helped |
| Apply the proposed method rather than WGANGP | MLP generator · MNIST training | FID: 8.46 -> 3.21 | Helped |
| Apply the proposed method rather than WGANGP | DCGAN generator · MNIST test | NND-WC: 0.07 -> 0.09 | Hurt |
| Apply the proposed method rather than WAE | DCGAN generator · Thin-8 | WD: 1950 -> 923 | Helped |
| Apply the proposed method rather than WGANGP | DCGAN generator · Thin-8 | NND-GP: 79.47 -> 47.59 | Helped |
| Minimize pixel-wise Wasserstein distance rather than use an adversarial objective | Generative models · CIFAR10 | visual quality | Hurt |
| Use the proposed method rather than WGAN-GP | Generative models · CIFAR10 | WD: 849 -> 655 | Helped |
Closest papers by meaning
- Asymptotic Guarantees for Learning Generative Models with the Sliced-Wasserstein Distance (NeurIPS 2019 · similarity 0.90)
- Score-based Generative Modeling Secretly Minimizes the Wasserstein Distance (NeurIPS 2022 · similarity 0.90)
- Statistically Optimal Generative Modeling with Maximum Deviation from the Empirical Distribution (ICML 2024 · similarity 0.89)
- Scalable Wasserstein Gradient Flow for Generative Modeling through Unbalanced Optimal Transport (ICML 2024 · similarity 0.88)
- Generative Conditional Distributions by Neural (Entropic) Optimal Transport (ICML 2024 · similarity 0.87)
- Towards Understanding Gradient Dynamics of the Sliced-Wasserstein Distance via Critical Point Analysis (ICML 2025 · 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.