Research map / Probabilistic and scientific modeling / Variational and generative sampling
Sequential Monte Carlo for inference
Uses twisted and learned sequential Monte Carlo for language-model and latent-variable inference.
78 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 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.
16 papers in 2023–24 and 18 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic held steady, from 3.9% 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
- 2016 · Inference Networks for Sequential Monte Carlo in Graphical Models (ICML 2016 · 48 citations)
- 2017 · Learning Deep Latent Gaussian Models with Markov Chain Monte Carlo (ICML 2017 · 57 citations)
- 2018 · Auto-Encoding Sequential Monte Carlo (ICLR 2018 · 71 citations)
- 2019 · Divergence Triangle for Joint Training of Generator Model, Energy-Based Model, and Inferential Model (CVPR 2019 · 52 citations)
- 2020 · On the Anatomy of MCMC-Based Maximum Likelihood Learning of Energy-Based Models (AAAI 2020 · 72 citations)
- 2021 · Learning Generative Vision Transformer with Energy-Based Latent Space for Saliency Prediction (NeurIPS 2021 · 45 citations)
- 2022 · Exposing the Implicit Energy Networks behind Masked Language Models via Metropolis--Hastings (ICLR 2022 · 7 citations)
- 2023 · Sampling-Based Accuracy Testing of Posterior Estimators for General Inference (ICML 2023 · 18 citations)
- 2024 · Using Stratified Sampling to Improve LIME Image Explanations (AAAI 2024 · 7 citations)
- 2025 · Fast training and sampling of Restricted Boltzmann Machines (ICLR 2025 · 2 citations)
- 2026 · A Probabilistic Framework for LLM-Based Model Discovery (ICML 2026 · 0 citations)
Earliest papers here
- Inference Networks for Sequential Monte Carlo in Graphical Models (ICML 2016 · 48 citations)
- Neurally-Guided Procedural Models: Amortized Inference for Procedural Graphics Programs using Neural Networks (NeurIPS 2016 · 32 citations)
- Pairwise Choice Markov Chains (NeurIPS 2016 · 28 citations)
What has been tried
67 papers here report 619 measured changes (a component added, removed or swapped, compared with a baseline): helped 68%, no clear effect 10%, hurt 15%, 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 |
|---|---|---|---|---|
| Add uniform samples to the soft-mining batch | LLFF, 1k iterations · PSNR | 23.91 -> 24.04 | Helped | Accelerating Neural Field Training via Soft Mining (CVPR 2024 · 7 citations) |
| Add low-rank RBM pretraining before PCD | Highly structured datasets · test log-likelihood | Helped | Fast training and sampling of Restricted Boltzmann Machines (ICLR 2025 · 2 citations) | |
| Increase the pCNL step size to 2.0 rather than using MALA at 0.05 | Layout-to-image generation · sample diversity (LPIPS MPD) | No clear effect | $\Psi$-Sampler: Initial Particle Sampling for SMC-Based Inference-Time Reward Alignment in Score Models (NeurIPS 2025 · 1 citation) | |
| Increase the number of decoding steps | WMT23 English→Russian · average reward | Helped | QUEST: Quality-Aware Metropolis-Hastings Sampling for Machine Translation (NeurIPS 2024 · 1 citation) | |
| train on fixed-length Δ-energy episodes rather than first/best episodes | ACE2 stability · proportion ddG < -5 | 0.609 -> 0.748 | Helped | Learning Extrapolative Sequence Transformations from Markov Chains (ICML 2025 · 0 citations) |
| Increase MCMC steps from k=2 to k=10 | SQL · branch coverage | 29.47(0.75) -> 32.93(0.83) | Helped | Constrained Sampling for Language Models Should Be Easy: An MCMC Perspective (NeurIPS 2025 · 0 citations) |
| add a grammar constraint | Planetarium goal inference · accuracy | 0.063 -> 0.086 | Helped | Syntactic and Semantic Control of Large Language Models via Sequential Monte Carlo (ICLR 2025 · 0 citations) |
| Add length normalization to model probabilities | WMT'19 De-En and Ru-En · BLEU | #wins MBMBR\textsubscript{L}{} > MBR{} is 47/56 | Helped | Model-Based Minimum Bayes Risk Decoding for Text Generation (ICML 2024 · 0 citations) |
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 34 papers; from OpenAlex and the papers' affiliations).
- University of British Columbia · 3 papers
- Google · 2 papers
- University of Toronto · 2 papers
- Simon Fraser University · 2 papers
- Johns Hopkins University · 2 papers
- ETHZ - ETH Zurich · 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
- Hamiltonian Monte Carlo (89 papers)
- Stochastic gradient MCMC (68 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).