Research map / Probabilistic and scientific modeling / Variational and generative sampling
Wasserstein Auto-Encoders
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 WAE-MMD instead of VAE | CelebA generative modeling · CelebA | FID: 63 -> 55 | Helped |
| Use WAE-GAN instead of VAE | CelebA generative modeling · CelebA | FID: 63 -> 42 | Helped |
| Use WAE-GAN instead of WAE-MMD | CelebA generative modeling · CelebA | FID: 55 -> 42 | Helped |
| Use WAE-MMD instead of VAE | CelebA generative modeling · CelebA | sharpness: $3\times10^{-3}$ -> $6\times10^{-3}$ | Helped |
| Use WAE-GAN instead of VAE | CelebA generative modeling · CelebA | sharpness: $3\times10^{-3}$ -> $6\times10^{-3}$ | Helped |
| Use bigWAE-MMD instead of bigVAE | CelebA large-scale model study · CelebA | FID: 45 -> 37 | Helped |
| Use bigWAE-GAN instead of bigVAE | CelebA large-scale model study · CelebA | FID: 45 -> 35 | Helped |
| Use bigWAE-GAN instead of bigWAE-MMD | CelebA large-scale model study · CelebA | FID: 37 -> 35 | Helped |
| Replace the RBF kernel with an inverse multiquadratics kernel (ablation) | WAE-MMD · MNIST and CelebA | sample quality | Helped |
| Use an RBF kernel for latent matching (ablation) | WAE-MMD · MNIST and CelebA | outlier penalization | Hurt |
| Combine encoder pretraining with truncated pixel-wise Gaussian input noise (ablation) | MNIST WAE · MNIST | results | Helped |
| Use adversarial latent matching rather than MMD matching | WAE · CelebA | training stability | Hurt |
Closest papers by meaning
- Statistical Regeneration Guarantees of the Wasserstein Autoencoder with Latent Space Consistency (NeurIPS 2021 · similarity 0.92)
- A Statistical Analysis of Wasserstein Autoencoders for Intrinsically Low-dimensional Data (ICLR 2024 · similarity 0.90)
- A Primal-Dual link between GANs and Autoencoders (NeurIPS 2019 · similarity 0.90)
- Sliced Wasserstein Auto-Encoders (ICLR 2019 · similarity 0.90)
- Gromov-Wasserstein Autoencoders (ICLR 2023 · similarity 0.88)
- StrWAEs to Invariant Representations (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.