Research map / Learning theory and algorithms / Optimal transport
COT-GAN: Generating Sequential Data via Causal Optimal Transport
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 COT-GAN's causal transport objective rather than MoCoGAN | video generation · Sprites | FVD: 1\,108.2 -> 458.0 | Helped |
| use COT-GAN's causal transport objective rather than direct mixed Sinkhorn minimization | video generation · Sprites | FID: 81.56 -> 84.6 | Hurt |
| use COT-GAN's causal transport objective rather than MoCoGAN | video generation · Human actions | FID: 151.3 -> 58.9 | Helped |
| use COT-GAN's causal transport objective rather than direct mixed Sinkhorn minimization | video generation · Human actions | KVD: 34.3 -> 43.7 | Hurt |
| replace the mixed Sinkhorn divergence with the six-term mini-batch Sinkhorn divergence (ablation) | COT-GAN · Sprites | FVD: 458.0 -> 620.1 | Hurt |
| replace the mixed Sinkhorn divergence with the six-term mini-batch Sinkhorn divergence (ablation) | COT-GAN · Sprites | KVD: 66.1 -> 64.5 | Helped |
| replace the mixed Sinkhorn divergence with the six-term mini-batch Sinkhorn divergence (ablation) | COT-GAN · Human actions | FVD: 462.8 -> 630.8 | Hurt |
| replace the mixed Sinkhorn divergence with the six-term mini-batch Sinkhorn divergence (ablation) | COT-GAN · Human actions | KVD: 43.7 -> 46.79 | Hurt |
| use the mixed Sinkhorn divergence instead of the original Sinkhorn divergence (ablation) | COT-GAN on multivariate AR-1 · multivariate AR-1 | correlation matching | Helped |
| add the martingale penalty (ablation) | COT-GAN on multivariate AR-1 · multivariate AR-1 | correlation matching | Helped |
Closest papers by meaning
- Parallel Optimal Transport GAN (CVPR 2019 · similarity 0.89)
- CoT: Cooperative Training for Generative Modeling of Discrete Data (ICML 2019 · similarity 0.89)
- Improving GANs Using Optimal Transport (ICLR 2018 · similarity 0.89)
- Analyzing and Improving Optimal-Transport-based Adversarial Networks (ICLR 2024 · similarity 0.88)
- MCGAN: Enhancing GAN Training with Regression-Based Generator Loss (AAAI 2025 · similarity 0.88)
- On the Convergence and Robustness of Training GANs with Regularized Optimal Transport (NeurIPS 2018 · similarity 0.88)
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.