Research map / Probabilistic and scientific modeling / Variational and generative sampling
Bayes-Factor-VAE: Hierarchical Bayesian Deep Auto-Encoder Models for Factor Disentanglement
Minyoung Kim, Yuting Wang, Pritish Sahu, Vladimir Pavlovic
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 |
|---|---|---|---|
| use BF-VAE-2 hierarchical prior with relevance indicators instead of F-VAE | Factor-VAE · Sprites | Metric II accuracy: 77.6 -> 85.9 | Improved |
| use BF-VAE-2 prior instead of F-VAE | Factor-VAE · O-Spr | Metric II accuracy: 80.6 -> 97.1 | Improved |
| use BF-VAE-1 prior instead of F-VAE | Factor-VAE · Teapots | Metric II accuracy: 74.6 -> 82.7 | Improved |
| use BF-VAE-1 hierarchical Gamma prior instead of F-VAE | Factor-VAE · 3D-Face | Metric II accuracy: 92.8 -> 97.2 | Improved |
| use BF-VAE-1 instead of RF-VAE (relevance without explicit non-Gaussianity) | RF-VAE · Sprites | Metric I accuracy: 85.4 -> 93.8 | Improved |
| use BF-VAE-2 instead of RF-VAE | RF-VAE · C-Spr | Metric I accuracy: 80.7 -> 94.7 | Improved |
| use BF-VAE-1 or BF-VAE-2 instead of RF-VAE | RF-VAE · Teapots | Metric I accuracy: 98.7 vs 97.6 and 97.9 | No clear effect |
| use BF-VAE-0 instead of BF-VAE-1 | BF-VAE · Teapots | Metric II accuracy: 82.7 -> 90.4 | Improved |
| replace the Gaussian prior with a high-capacity Gaussian mixture prior in F-VAE (ablation) | Factor-VAE · O-Spr and Sprites | Metric II accuracy | Hurt |
| increase mixture components K of the MoG prior (ablation) | Factor-VAE with MoG prior · O-Spr and Sprites | Metric II accuracy | Hurt |
| increase the trade-off parameter eta (more weight on sparsity of relevance) (ablation) | BF-VAE-2 · Celeb-A | number of detected relevant factors | Improved |
Closest papers by meaning
- Disentangling Disentanglement in Variational Autoencoders (ICML 2019 · similarity 0.91)
- Hyperprior Induced Unsupervised Disentanglement of Latent Representations (AAAI 2019 · similarity 0.90)
- Learning Discrete and Continuous Factors of Data via Alternating Disentanglement (ICML 2019 · similarity 0.89)
- Demystifying Inductive Biases for (Beta-)VAE Based Architectures (ICML 2021 · similarity 0.89)
- $\alpha$TC-VAE: On the relationship between Disentanglement and Diversity (ICLR 2024 · similarity 0.88)
- Disentanglement Beyond Static vs. Dynamic: A Benchmark and Evaluation Framework for Multi-Factor Sequential Representations (NeurIPS 2025 · 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.
