Research map / Probabilistic and scientific modeling / Variational and generative sampling
Disentangled Recurrent Wasserstein Autoencoder
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 R-WAE(MMD) instead of DS-VAE with the same setting | Sequential disentangled representation learning · TIMIT (speaker, Zc) | EER: 5.64% -> 4.73% | Helped |
| Use R-WAE(MMD) instead of DS-VAE(S) | Sprites simple architecture · Sprites actions | classification error: 8.11% -> 5.83% | Helped |
| Use R-WAE(MMD) instead of DS-VAE(S) | Sequential disentangled representation learning · SM-MNIST digits | classification error: 3.31% -> 1.78% | Helped |
| Use a complex rather than simple encoder/decoder architecture (ablation) | R-WAE(MMD) · Sprites actions | classification error: 5.83% -> 7.72% | Hurt |
| Use a complex rather than simple encoder/decoder architecture (ablation) | R-WAE(MMD) · Sprites content | classification error: 2.45% -> 3.31% | Hurt |
| Add the known number of actions as categorical latent prior information (ablation) | DS-VAE on MUG · MUG generated samples | Accuracy: 66.73% -> 82.84 % | Helped |
| Use R-WAE(GAN) rather than R-WAE(MMD) for content regularization (ablation) | R-WAE on MUG · MUG generated samples | FID: 12.21 -> 10.86 | Helped |
| Use R-WAE(GAN) rather than R-WAE(MMD) for content regularization (ablation) | R-WAE on MUG · MUG generated samples | Accuracy: 88.62% -> 90.15% | Helped |
Closest papers by meaning
- S3VAE: Self-Supervised Sequential VAE for Representation Disentanglement and Data Generation (CVPR 2020 · similarity 0.88)
- Contrastively Disentangled Sequential Variational Autoencoder (NeurIPS 2021 · similarity 0.88)
- Sequential Disentanglement by Extracting Static Information From A Single Sequence Element (ICML 2024 · similarity 0.87)
- A Statistical Analysis of Wasserstein Autoencoders for Intrinsically Low-dimensional Data (ICLR 2024 · similarity 0.87)
- StrWAEs to Invariant Representations (ICML 2024 · similarity 0.86)
- Transferring disentangled representations: bridging the gap between synthetic and real images (NeurIPS 2024 · 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.