Research map / Probabilistic and scientific modeling / Variational and generative sampling
Disentangled and discrete latent models
Learns discrete or disentangled latent concepts with identifiability analysis.
92 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2017–2026) · an approach in Variational and generative sampling, keyword cluster “vae · autoencoders · latent”
State History Tried Checklist Preprints Who Export
Where it stands
Not enough benchmark data Fewer than four papers report the same benchmark, so there is no record to follow.
10 papers in 2023–24 and 18 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic grew, from 2.4% to 3.6%.
How this is decided
A benchmark is tracked when at least four papers of this approach report their own method on the same dataset and metric in at least two different years. Its record is the best number reported up to 2024 compared with the best reported in 2025–26: still improving means at least 1% better, flat means at most 0.2% better with at least two papers from 2025–26 trying. Numbers on a different scale from the rest (a fraction among percentages) and far outliers are left out. Results come from different papers with different setups, so they are signals, not head-to-head comparisons. The share of papers is counted at ICML, NeurIPS, ICLR, CVPR and AAAI.
How it got here
Most cited paper of each year
- 2017 · beta-VAE: Learning Basic Visual Concepts with a Constrained Variational Framework (ICLR 2017 · 2,979 citations)
- 2018 · Cooperative Learning of Energy-Based Model and Latent Variable Model via MCMC Teaching (AAAI 2018 · 59 citations)
- 2019 · Deep Latent Generative Models for Energy Disaggregation (AAAI 2019 · 39 citations)
- 2020 · UC-Net: Uncertainty Inspired RGB-D Saliency Detection via Conditional Variational Autoencoders (CVPR 2020 · 412 citations)
- 2021 · Disentangled Recurrent Wasserstein Autoencoder (ICLR 2021 · 13 citations)
- 2022 · Glass: Geometric Latent Augmentation for Shape Spaces (CVPR 2022 · 12 citations)
- 2023 · A Unified Knowledge Distillation Framework for Deep Directed Graphical Models (CVPR 2023 · 1 citation)
- 2024 · MorphVAE: Advancing Morphological Design of Voxel-Based Soft Robots with Variational Autoencoders (AAAI 2024 · 6 citations)
- 2025 · Generalization Bounds via Meta-Learned Model Representations: PAC-Bayes and Sample Compression Hypernetworks (ICML 2025 · 0 citations)
- 2026 · Unsupervised Disentanglement Without Compromises : How Functional Orthogonality Enforces Identifiability (ICML 2026 · 0 citations)
Earliest papers here
- beta-VAE: Learning Basic Visual Concepts with a Constrained Variational Framework (ICLR 2017 · 2,979 citations)
- A Hierarchical Latent Variable Encoder-Decoder Model for Generating Dialogues (AAAI 2017 · 701 citations)
- Adversarially Learned Inference (ICLR 2017 · 663 citations)
What has been tried
74 papers here report 714 measured changes (a component added, removed or swapped, compared with a baseline): helped 68%, no clear effect 6%, hurt 21%, mixed 5%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Use MorphVAE-B instead of MorphVAE-H, with lower selection temperature and exploration-exploitation rebalancing | Six Evolution Gym tasks · robot diversity | 2.17 average rank at k=10; 1.83 average rank at k=5, compared with 5.83 and 5.17 for MorphVAE-H | Helped | MorphVAE: Advancing Morphological Design of Voxel-Based Soft Robots with Variational Autoencoders (AAAI 2024 · 6 citations) |
| Disable denoising noise during training (σ=0) | MIT multi-illumination test · SSIM (raw) | 0.473 -> 0.232 | Hurt | Latent Intrinsics Emerge from Training to Relight (NeurIPS 2024 · 5 citations) |
| use a learned binary message in addition to the compression set (SCH+ vs SCH-) | MNIST-pixels-swap, 100 Pixels swap · Bound | 0.067 ± 0.007 -> 0.035 ± 0.012 | Helped | Generalization Bounds via Meta-Learned Model Representations: PAC-Bayes and Sample Compression Hypernetworks (ICML 2025 · 0 citations) |
| remove equivariance constraints from the ENF | CIFAR10 test · accuracy | 82.1 -> 68.7 | Hurt | Grounding Continuous Representations in Geometry: Equivariant Neural Fields (ICLR 2025 · 0 citations) |
| Remove the sparse mixing constraint L_s | synthetic datasets with fewer auxiliary domains · MCC | Hurt | Synergy Between Sufficient Changes and Sparse Mixing Procedure for Disentangled Representation Learning (ICLR 2025 · 0 citations) | |
| set α to 0.5 rather than α=1 | TIMIT inference · phoneme duration MAE | 3.93 ± 0.37 -> 4.40 ± 0.46 | Hurt | Latent Optimal Paths by Gumbel Propagation for Variational Bayesian Dynamic Programming (ICML 2024 · 0 citations) |
| increase context size from M=32 to M=128 | FER2013, CelebA, CLEVR, t-ImageNet · log marginal likelihood | 4111 -> 4951; 11569 -> 14210; 16529 -> 16611; 15645 -> 16324 | Helped | Autoencoding Conditional Neural Processes for Representation Learning (ICML 2024 · 0 citations) |
What reviewers will expect
Datasets and baselines named in the 6 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- MNIST · 2 of 6 papers
- CelebA · 2 of 6 papers
Baselines
- beta-VAE · 2 of 6 papers
- iVAE · 2 of 6 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 28 papers; from OpenAlex and the papers' affiliations).
- University of California, Los Angeles · 2 papers
- University of Edinburgh · 2 papers
Use it
Download BibTeX of 40 key papersOpen the topic in the interactive map
Get alerts. Describe your idea in scime atlas and save it (free scime account): you get a weekly email when new preprints land close to it.
Related approaches
Same keyword cluster
- Structured latent VAEs (125 papers)
- Regularizing VAE latent information (89 papers)
- Hierarchical VAEs and posterior collapse (59 papers)
Elsewhere in the topic
- Likelihood-free and graphical model estimation (161 papers)
- Flow matching and flow sampling (144 papers)
- Posterior sampling with diffusion guidance (137 papers)
- Score-based generative modeling analysis (128 papers)
- Exact inference for probabilistic programs (103 papers)
- Latent variable marginalization and estimators (91 papers)
- Tractable probabilistic models (90 papers)
- Variational inference theory and gradients (89 papers)
Benchmark numbers, ablations and baseline numbers are extracted by AI from each paper's full text and can be wrong; the paper is the authority. Data: results index of 2026-11-02, citation counts from OpenAlex (2026-10-09).