Research map / Probabilistic and scientific modeling / Variational and generative sampling
Cramer-Wold Auto-Encoder
Szymon Knop, Przemysław Spurek, Jacek Tabor, Igor Podolak, Marcin Mazur, Stanisław Jastrzębski
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 |
|---|---|---|---|
| add logarithm to the distance term of the SWAE cost (ablation) | SWAE · MNIST | FID: 29.76 -> 24.89 | Improved |
| add logarithm to the distance term of the SWAE cost (ablation) | SWAE · Fashion-MNIST | FID: 74.32 -> 61.72 | Improved |
| add logarithm to the distance term of the SWAE cost (ablation) | SWAE · CIFAR-10 | FID: 141.91 -> 119.89 | Improved |
| add logarithm to the distance term of the SWAE cost (ablation) | SWAE · CelebA | FID: 53.85 -> 53.46 | No clear effect |
| add logarithm to the MMD distance term of WAE-MMD (ablation) | WAE-MMD · MNIST | FID: 27.65 -> 36.08 | Hurt |
| add logarithm to the MMD distance term of WAE-MMD (ablation) | WAE-MMD · Fashion-MNIST | FID: 69.16 -> 99.51 | Hurt |
| add logarithm to the MMD distance term of WAE-MMD (ablation) | WAE-MMD · CIFAR-10 | FID: 129.37 -> 136.25 | Hurt |
| add logarithm to the MMD distance term of WAE-MMD (ablation) | WAE-MMD · CelebA | FID: 51.51 -> 58.10 | Hurt |
| replace WAE-MMD sampled MMD penalty with closed-form Cramer-Wold distance with log (CWAE) | WAE-MMD · MNIST | FID: 27.65 -> 23.63 | Improved |
| replace WAE-MMD sampled MMD penalty with closed-form Cramer-Wold distance with log (CWAE) | WAE-MMD · Fashion-MNIST | FID: 69.16 -> 57.06 | Improved |
| replace WAE-MMD sampled MMD penalty with closed-form Cramer-Wold distance with log (CWAE) | WAE-MMD · CIFAR-10 | FID: 129.37 -> 120.02 | Improved |
| replace WAE-MMD sampled MMD penalty with closed-form Cramer-Wold distance with log (CWAE) | WAE-MMD · CelebA | FID: 51.51 -> 49.69 | Improved |
Closest papers by meaning
- Unscented Autoencoder (ICML 2023 · similarity 0.86)
- Model Selection for Bayesian Autoencoders (NeurIPS 2021 · similarity 0.85)
- A Statistical Analysis of Wasserstein Autoencoders for Intrinsically Low-dimensional Data (ICLR 2024 · similarity 0.85)
- Gromov-Wasserstein Autoencoders (ICLR 2023 · similarity 0.85)
- Statistical Regeneration Guarantees of the Wasserstein Autoencoder with Latent Space Consistency (NeurIPS 2021 · similarity 0.85)
- StrWAEs to Invariant Representations (ICML 2024 · similarity 0.84)
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.
