Research map / Image generation / Autoregressive and latent generation
An Architecture for Deep, Hierarchical Generative Models
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 lightweight local autoregressive model to reconstruction distribution | MatNet on CIFAR 10 · CIFAR 10 | bits per pixel | Helped |
| Gaussian reconstruction without autoregression on CIFAR 10 | MatNet · CIFAR 10 | bits per pixel | No clear effect |
| place GMM prior over top latent z0 with no label information | MatNet on Omniglot · Omniglot | qualitative class structure | Helped |
| regularize inference model by maximizing free-energy on model samples (Eqn. 13) | MatNet on MNIST and Omniglot · MNIST, Omniglot | test NLL | Helped |
Closest papers by meaning
- Boosting Generative Image Modeling via Joint Image-Feature Synthesis (NeurIPS 2025 · similarity 0.90)
- Learning Hierarchical Features from Deep Generative Models (ICML 2017 · similarity 0.90)
- Flexible and accurate inference and learning for deep generative models (NeurIPS 2018 · similarity 0.89)
- Auxiliary Deep Generative Models (ICML 2016 · similarity 0.89)
- Masked Autoencoders Are Effective Tokenizers for Diffusion Models (ICML 2025 · similarity 0.87)
- Nested Diffusion Models Using Hierarchical Latent Priors (CVPR 2025 · 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.