Research map / Probabilistic and scientific modeling / Variational and generative sampling
Stochastic gradient MCMC
Samples large-scale distributions with stochastic gradient and pseudo-extended samplers.
68 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2026) · an approach in Variational and generative sampling, keyword cluster “monte · carlo · mcmc”
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.
11 papers in 2023–24 and 11 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 2.7% to 2.2%.
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 · Ensuring Rapid Mixing and Low Bias for Asynchronous Gibbs Sampling (ICML 2016 · 20 citations)
- 2017 · Measuring Sample Quality with Kernels (ICML 2017 · 81 citations)
- 2018 · Beyond Log-concavity: Provable Guarantees for Sampling Multi-modal Distributions using Simulated Tempering Langevin Monte Carlo (NeurIPS 2018 · 16 citations)
- 2019 · Probabilistic Planning with Sequential Monte Carlo methods (ICLR 2019 · 27 citations)
- 2020 · Hamiltonian Monte Carlo using an adjoint-differentiated Laplace approximation: Bayesian inference for latent Gaussian models and beyond (NeurIPS 2020 · 9 citations)
- 2021 · Bayesian Optimized Monte Carlo Planning (AAAI 2021 · 20 citations)
- 2022 · Variational inference via Wasserstein gradient flows (NeurIPS 2022 · 12 citations)
- 2023 · Wasserstein Quantum Monte Carlo: A Novel Approach for Solving the Quantum Many-Body Schrödinger Equation (NeurIPS 2023 · 3 citations)
- 2024 · Markovian Flow Matching: Accelerating MCMC with Continuous Normalizing Flows (NeurIPS 2024 · 3 citations)
- 2025 · Multilevel Generative Samplers for Investigating Critical Phenomena (ICLR 2025 · 1 citation)
- 2026 · Can Microcanonical Langevin Dynamics Leverage Mini-Batch Gradient Noise? (ICML 2026 · 0 citations)
Earliest papers here
- Ensuring Rapid Mixing and Low Bias for Asynchronous Gibbs Sampling (ICML 2016 · 20 citations)
- Measuring the reliability of MCMC inference with bidirectional Monte Carlo (NeurIPS 2016 · 16 citations)
- Interacting Particle Markov Chain Monte Carlo (ICML 2016 · 10 citations)
What has been tried
59 papers here report 539 measured changes (a component added, removed or swapped, compared with a baseline): helped 67%, no clear effect 11%, hurt 15%, mixed 7%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| change k_Q from K to 10^2 | 4-mode Gaussian mixture · MMD | 2.37e-3 ± 2.29e-3 -> 1.05e-3 ± 8.90e-4 | Helped | Markovian Flow Matching: Accelerating MCMC with Continuous Normalizing Flows (NeurIPS 2024 · 3 citations) |
| use sequential training with model transfer instead of random initialization | 2D Ising, N=64 · training efficiency | Helped | Multilevel Generative Samplers for Investigating Critical Phenomena (ICLR 2025 · 1 citation) | |
| set power parameter p to 2 instead of 1 | CartPole-IG · reward | 76.91 -> 77.85 | Helped | Power Mean Estimation in Stochastic Continuous Monte-Carlo Tree Search (ICML 2025 · 0 citations) |
| use symmetric thresholds on the absolute log acceptance ratio instead of the asymmetric criterion | Normal(0,1), Laplace(0,1), and Cauchy(0,1), across state norms 10^{-5} to 10^2 · acceptance probability | Helped | AutoStep: Locally adaptive involutive MCMC (ICML 2025 · 0 citations) | |
| Estimate the Jacobian diagonal stochastically instead of using automatic differentiation | German Credit Dataset · wall-clock time | over 10\times faster | Helped | Parallelizing MCMC Across the Sequence Length (NeurIPS 2025 · 0 citations) |
| use Langevin noise instead of trajectory length randomization | Gaussian · gradient calls to squared error < 0.01 | 3,249 -> 3,172 | Helped | Metropolis Adjusted Microcanonical Hamiltonian Monte Carlo (NeurIPS 2025 · 0 citations) |
| Add expanded parameterization | CIFAR-10 test · ERR | 0.135\PM{0.003} -> 0.115\PM{0.002} | Helped | Parameter Expanded Stochastic Gradient Markov Chain Monte Carlo (ICLR 2025 · 0 citations) |
| Use BMA meta-loss instead of cross-entropy meta-loss | CIFAR-10 · loss barrier between collected parameters | Helped | Learning to Explore for Stochastic Gradient MCMC (ICML 2024 · 0 citations) | |
| Add boundary integral estimation | Ring, Cardioid, Double-moon, and Block · Wasserstein-2 distance (Sinkhorn) | 0.2138 -> 0.1087; 0.2321 -> 0.1660; 0.4866 -> 0.1141; 0.2438 -> 0.2416 | Helped | Functional Gradient Flows for Constrained Sampling (NeurIPS 2024 · 0 citations) |
| Marginalize u2 rather than leave random effects untransformed | Grouse ticks · divergent transitions | 42.60 (25.76) -> 0.00 (0.00) | Helped | Hamiltonian Monte Carlo Inference of Marginalized Linear Mixed-Effects Models (NeurIPS 2024 · 0 citations) |
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.
Baselines
- NUTS · 3 of 7 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 22 papers; from OpenAlex and the papers' affiliations).
- University of Massachusetts Amherst · 2 papers
- Stanford University · 2 papers
- Peking University · 2 papers
Use it
Download BibTeX of 36 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
- Hamiltonian Monte Carlo (89 papers)
- Sequential Monte Carlo for inference (78 papers)
- Importance sampling and quadrature (45 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).