Research map / Probabilistic and scientific modeling / Variational and generative sampling
Variational Autoencoders with Riemannian Brownian Motion Priors
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 a Riemannian Brownian motion prior instead of the standard Gaussian prior | VAE · MNIST, d=2 | average F1: 0.72 ±.002 -> 0.78 ±.002 | Helped |
| Use a Riemannian Brownian motion prior instead of the standard Gaussian prior | VAE · MNIST, d=5 | average F1: 0.92 ±.001 -> 0.93 ±.0008 | Helped |
| Use a Riemannian Brownian motion prior instead of the standard Gaussian prior | VAE · MNIST, d=10 | average F1: 0.96 ±.001 -> 0.96 ±.001 | No clear effect |
| Use a Riemannian Brownian motion prior | VAE · MNIST | Neg. ELBO: -1030.38 ±5.34 -> -1047.29 ±2.77; -1076.64 ±4.48 -> -1141.06 ±7.09; -1110.79 ±1.17 -> -1170.03 ±18.52 | Mixed |
| Use a Riemannian Brownian motion prior | VAE · FashionMNIST | Neg. ELBO: -443.13 ±10.67 -> -708.77 ±6.93; -511.65 ±3.70 -> -889.62 ±3.44; -525.05 ±5.87 -> -959.2 ±5.37 | Mixed |
| Use a Riemannian Brownian motion prior instead of the standard Gaussian prior | VAE · MNIST | training time per epoch: 10.64 ±.51 -> 55.73 ±4.36 | Hurt |
Closest papers by meaning
- A Geometric Perspective on Variational Autoencoders (NeurIPS 2022 · similarity 0.92)
- Variational Autoencoder with Implicit Optimal Priors (AAAI 2019 · similarity 0.89)
- Constraining Variational Inference with Geometric Jensen-Shannon Divergence (NeurIPS 2020 · similarity 0.89)
- Mixed-curvature Variational Autoencoders (ICLR 2020 · similarity 0.89)
- Gaussian Process Prior Variational Autoencoders (NeurIPS 2018 · similarity 0.89)
- $t^3$-Variational Autoencoder: Learning Heavy-tailed Data with Student's t and Power Divergence (ICLR 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.