Research map / Probabilistic and scientific modeling / Variational and generative sampling
Coupled Variational Autoencoder
Xiaoran Hao, Patrick Shafto
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 |
|---|---|---|---|
| Recast VAE training as entropic OT between prior and data with hard aggregate-posterior matching (C-VAE) | VAE on mixture of Gaussians · Mixture of 25 Gaussians | High density ratio: 98.5 vs 86.0 for InfoVAE, 71.5 for VAE | Improved |
| Recast VAE training as entropic OT with hard prior-matching constraint | VAE on mixture of Gaussians · Mixture of 25 Gaussians | MMD between prior and aggregate posterior: 0.0032 vs 0.0045 for InfoVAE, 0.0106 for VAE | Improved |
| Replace KL-based regularizer with EOT coupling (C-VAE) vs InfoVAE penalty | InfoVAE-style prior matching on mixture of Gaussians · Mixture of 25 Gaussians | std within modes (ground truth 0.05): 0.0478 vs 0.0673 for InfoVAE | Improved |
| Normalizing flow posterior (VAE-NF) instead of Gaussian posterior | VAE · Mixture of 25 Gaussians | High density ratio: 53.6 vs 71.5 for VAE | Hurt |
| Normalizing flow posterior (VAE-NF) instead of Gaussian posterior | VAE · Mixture of 25 Gaussians | MMD prior vs aggregate posterior: 0.0161 vs 0.0106 for VAE | Hurt |
Closest papers by meaning
- Variational Autoencoders Pursue PCA Directions (by Accident) (CVPR 2019 · similarity 0.90)
- From Variational to Deterministic Autoencoders (ICLR 2020 · similarity 0.90)
- Adversarial Symmetric Variational Autoencoder (NeurIPS 2017 · similarity 0.89)
- Connections with Robust PCA and the Role of Emergent Sparsity in Variational Autoencoder Models (JMLR 2018 · similarity 0.89)
- Variational Autoencoder with Implicit Optimal Priors (AAAI 2019 · similarity 0.89)
- Generalization in VAE and Diffusion Models: A Unified Information-Theoretic Analysis (ICLR 2025 · similarity 0.89)
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.
