Research map / Probabilistic and scientific modeling / Variational and generative sampling
A Geometric Perspective on Variational Autoencoders
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 Riemannian-uniform sampling instead of the standard Gaussian prior | Vanilla VAE · MNIST | FID: 40.70 -> 8.53 | Helped |
| Use Riemannian-uniform sampling instead of the standard Gaussian prior | Vanilla VAE · SVHN | FID: 83.55 -> 46.99 | Helped |
| Use Riemannian-uniform sampling instead of the standard Gaussian prior | Vanilla VAE · CIFAR 10 | FID: 162.58 -> 93.53 | Helped |
| Use Riemannian-uniform sampling instead of the standard Gaussian prior | Vanilla VAE · CELEBA | FID: 64.13 -> 48.71 | Helped |
| Use Riemannian-uniform sampling instead of standard-prior sampling | VAE · OASIS | balanced accuracy: 51.8 ± 3.8 -> 85.4 ± 2.5 | Helped |
| Use Riemannian-uniform sampling rather than a 10-component latent GMM | VAE · MNIST | FID: 13.13 -> 8.53 | Helped |
| Use Riemannian-uniform sampling rather than a 10-component latent GMM | VAE · SVHN | FID: 52.32 -> 46.99 | Helped |
| Use Riemannian-uniform sampling rather than a 10-component latent GMM | VAE · CIFAR 10 | FID: 138.25 -> 93.53 | Helped |
| Use Riemannian-uniform sampling rather than a 10-component latent GMM | VAE · CELEBA | FID: 55.50 -> 48.71 | Helped |
| Use Riemannian-uniform sampling rather than prior sampling | Vanilla VAE · Low-data training | FID | Helped |
Closest papers by meaning
- Mixed-curvature Variational Autoencoders (ICLR 2020 · similarity 0.93)
- Variational Autoencoders with Riemannian Brownian Motion Priors (ICML 2020 · similarity 0.92)
- Continuous Hierarchical Representations with Poincaré Variational Auto-Encoders (NeurIPS 2019 · similarity 0.91)
- From Variational to Deterministic Autoencoders (ICLR 2020 · similarity 0.91)
- Constraining Variational Inference with Geometric Jensen-Shannon Divergence (NeurIPS 2020 · similarity 0.90)
- Graph Geometry-Preserving Autoencoders (ICML 2024 · similarity 0.88)
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.