Research map / Probabilistic and scientific modeling / Variational and generative sampling
Contrastively Disentangled Sequential Variational 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 |
|---|---|---|---|
| add content augmentation (ablation) | C-DSVAE · SM-MNIST | accuracy (%): 92.02 -> 97.84 | Helped |
| add motion augmentation (ablation) | C-DSVAE · SM-MNIST | accuracy (%): 95.25 -> 97.84 | Helped |
| add content augmentation (ablation) | C-DSVAE · TIMIT | motion EER (%): 24.30 -> 31.81 | Helped |
| add motion augmentation (ablation) | C-DSVAE · TIMIT | content EER (%): 4.31 -> 4.03 | Helped |
| use contrastive MI estimation instead of minibatch weighted sampling (ablation) | C-DSVAE · MUG | accuracy (%): 66.25 -> 81.16 | Helped |
| use contrastive MI estimation instead of minibatch weighted sampling (ablation) | C-DSVAE · SM-MNIST | accuracy (%): 91.81 -> 97.84 | Helped |
| increase batch size from 64 to 256 (ablation) | C-DSVAE · TIMIT | content EER (%): 4.21 -> 4.03 | Helped |
| C-DSVAE vs best prior (S3VAE/R-WAE) | C-DSVAE · MUG | accuracy (%): 71.25 -> 81.16 | Helped |
Closest papers by meaning
- S3VAE: Self-Supervised Sequential VAE for Representation Disentanglement and Data Generation (CVPR 2020 · similarity 0.91)
- Sample and Predict Your Latent: Modality-free Sequential Disentanglement via Contrastive Estimation (ICML 2023 · similarity 0.90)
- Multifactor Sequential Disentanglement via Structured Koopman Autoencoders (ICLR 2023 · similarity 0.90)
- Sequential Disentanglement by Extracting Static Information From A Single Sequence Element (ICML 2024 · similarity 0.89)
- $\alpha$TC-VAE: On the relationship between Disentanglement and Diversity (ICLR 2024 · similarity 0.86)
- Synergy Between Sufficient Changes and Sparse Mixing Procedure for Disentangled Representation Learning (ICLR 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.