Research map / Probabilistic and scientific modeling / Variational and generative sampling
Annealing and divergence-based variational objectives
Uses annealed importance sampling, alpha-divergences and thermodynamic objectives for variational bounds.
79 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2026) · an approach in Variational and generative sampling, keyword cluster “variational inference · vi · lower”
State History Tried Checklist Preprints Export
Where it stands
Not enough benchmark data Fewer than four papers report the same benchmark, so there is no record to follow.
13 papers in 2023–24 and 13 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 3.1% to 2.6%.
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
- 2016 · A Weighted Variational Model for Simultaneous Reflectance and Illumination Estimation (CVPR 2016 · 1,104 citations)
- 2017 · A Non-Convex Variational Approach to Photometric Stereo Under Inaccurate Lighting (CVPR 2017 · 56 citations)
- 2018 · Pathwise Derivatives Beyond the Reparameterization Trick (ICML 2018 · 60 citations)
- 2019 · A Variational EM Framework With Adaptive Edge Selection for Blind Motion Deblurring (CVPR 2019 · 42 citations)
- 2020 · Bi-level Score Matching for Learning Energy-based Latent Variable Models (NeurIPS 2020 · 7 citations)
- 2021 · Differentiable Particle Filtering via Entropy-Regularized Optimal Transport (ICML 2021 · 11 citations)
- 2022 · Escaping from the Barren Plateau via Gaussian Initializations in Deep Variational Quantum Circuits (NeurIPS 2022 · 22 citations)
- 2023 · Variational Wasserstein Barycenters with C-cyclical Monotonicity Regularization (AAAI 2023 · 3 citations)
- 2024 · GAD-PVI: A General Accelerated Dynamic-Weight Particle-Based Variational Inference Framework (AAAI 2024 · 5 citations)
- 2025 · Q-MAML: Quantum Model-Agnostic Meta-Learning for Variational Quantum Algorithms (AAAI 2025 · 7 citations)
- 2026 · Variational inference via Gaussian interacting particles in the Bures-Wasserstein geometry (ICML 2026 · 0 citations)
Earliest papers here
- A Weighted Variational Model for Simultaneous Reflectance and Illumination Estimation (CVPR 2016 · 1,104 citations)
- Relevant sparse codes with variational information bottleneck (NeurIPS 2016 · 60 citations)
- The Generalized Reparameterization Gradient (NeurIPS 2016 · 24 citations)
What has been tried
70 papers here report 615 measured changes (a component added, removed or swapped, compared with a baseline): helped 70%, no clear effect 10%, hurt 14%, mixed 6%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Use continuous weight adjustment instead of Duplicate/Kill | GP LIDAR, BLOB smoothing · W2 distance | 1.561e-01 ± 1.155e-03 -> 1.274e-01 ± 2.964e-04 | Helped | GAD-PVI: A General Accelerated Dynamic-Weight Particle-Based Variational Inference Framework (AAAI 2024 · 5 citations) |
| Replacing MSE with CS divergence prediction term in NIB, VIB-style baselines | ablation appendix · performance | Helped | Cauchy-Schwarz Divergence Information Bottleneck for Regression (ICLR 2024 · 2 citations) | |
| Use optimized approximation coefficients instead of piece-wise approximation coefficients | MA-fBM trajectory approximation · MSE | Helped | Variational Inference for SDEs Driven by Fractional Noise (ICLR 2024 · 2 citations) | |
| Use nuclear-norm scaling instead of Frobenius-norm scaling | Test JC across the reported datasets · Test JC | Mixed | Learning Set Functions with Implicit Differentiation (AAAI 2025 · 0 citations) | |
| use R2-G2 instead of RT in a one-layer VAE | MNIST · test ELBO | -94.22 ± 0.24 -> -94.39 ± 0.42 | Hurt | Rao-Blackwellised Reparameterisation Gradients (NeurIPS 2025 · 0 citations) |
| Set the control-variate coefficient c to 0 rather than a positive value | Variance experiment · estimator variance | No clear effect | Stochastic variance-reduced Gaussian variational inference on the Bures-Wasserstein manifold (ICLR 2025 · 0 citations) | |
| replace Poisson hidden spikes with categorical, then Gumbel-Softmax | Synthetic dataset · test log-likelihood | Hurt | A Differentiable Partially Observable Generalized Linear Model with Forward-Backward Message Passing (ICML 2024 · 0 citations) | |
| Increase DAIS0 transitions K | Bimodal Gaussian target · target-sample log density under q | Helped | Differentiable Annealed Importance Sampling Minimizes The Jensen-Shannon Divergence Between Initial and Target Distribution (ICML 2024) |
What reviewers will expect
Datasets and baselines named in the 7 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- CIFAR-10 · 2 of 7 papers
Recent preprints
Loading recent arXiv preprints…
Use it
Download BibTeX of 39 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
- Variational inference theory and gradients (89 papers)
- Amortized and online variational inference (59 papers)
- Variational methods for sequence models (40 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)
- Structured latent VAEs (125 papers)
- Exact inference for probabilistic programs (103 papers)
- Disentangled and discrete latent models (92 papers)
- Latent variable marginalization and estimators (91 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).