Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Nonparametric and PAC-Bayes Bayesian learning
Develops Bayesian nonparametric models and links PAC-Bayes theory with Bayesian inference.
130 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2026) · an approach in Uncertainty and Gaussian processes, keyword cluster “bayesian inference · posterior · priors”
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.
19 papers in 2023–24 and 38 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic grew, from 7.0% to 12.3%.
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 · Learning Treewidth-Bounded Bayesian Networks with Thousands of Variables (NeurIPS 2016 · 30 citations)
- 2017 · Personalizing Gesture Recognition Using Hierarchical Bayesian Neural Networks (CVPR 2017 · 22 citations)
- 2018 · An Interpretable Joint Graphical Model for Fact-Checking From Crowds (AAAI 2018 · 42 citations)
- 2019 · Bayesian Hierarchical Dynamic Model for Human Action Recognition (CVPR 2019 · 61 citations)
- 2020 · Learning under Model Misspecification: Applications to Variational and Ensemble methods (NeurIPS 2020 · 29 citations)
- 2021 · MARTA: Leveraging Human Rationales for Explainable Text Classification (AAAI 2021 · 42 citations)
- 2022 · Bayesian Invariant Risk Minimization (CVPR 2022 · 52 citations)
- 2023 · TabPFN: A Transformer That Solves Small Tabular Classification Problems in a Second (ICLR 2023 · 102 citations)
- 2024 · Bayesian Inference with Complex Knowledge Graph Evidence (AAAI 2024 · 3 citations)
- 2025 · Scalable Bayesian Learning with posteriors (ICLR 2025 · 3 citations)
- 2026 · Demystifying LLM-as-a-Judge: Analytically Tractable Model for Inference-Time Scaling (ICML 2026 · 0 citations)
Earliest papers here
- Learning Treewidth-Bounded Bayesian Networks with Thousands of Variables (NeurIPS 2016 · 30 citations)
- Estimating Accuracy from Unlabeled Data: A Bayesian Approach (ICML 2016 · 24 citations)
- Learning Bayesian networks with ancestral constraints (NeurIPS 2016 · 20 citations)
What has been tried
92 papers here report 809 measured changes (a component added, removed or swapped, compared with a baseline): helped 69%, no clear effect 10%, hurt 16%, 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 BIKG with the Marginal quantifier rather than MPE alone | Movielens 20M and LFM-1b critique-based recommendation · hit rate@10 | Helped | Bayesian Inference with Complex Knowledge Graph Evidence (AAAI 2024 · 3 citations) | |
| Increase consistency samples from K=5 to K=100 | High-dimensional time series (N=2048) · MMD | 0.78 ± 0.04 -> 0.54 ± 0.03 | Helped | Leveraging Self-Consistency for Data-Efficient Amortized Bayesian Inference (ICML 2024 · 0 citations) |
| increase recursion depth from T=2 to T=8 | MNIST · test 0-1 loss | .139 (3e-3) -> .097 (2e-3) | Helped | Recursive PAC-Bayes: A Frequentist Approach to Sequential Prior Updates with No Information Loss (NeurIPS 2024 · 0 citations) |
| Use PM2 instead of PM1 with a Fourier basis for μ | warpMix simulation, μ3 · estimation error | 0.0195 -> 0.0044 | Helped | Probabilistic size-and-shape functional mixed models (NeurIPS 2024 · 0 citations) |
| Increase VPT tree depth from L=4 to L=6 | POWER, GAS, HEPMASS, MINIBOONE, BSDS300 test sets · log-likelihood | 0.67 -> 0.61; 11.92 -> 12.20; -15.01 -> -13.94; -8.71 -> -8.51; 158.59 -> 162.72 | Mixed | Variational Polya Tree (NeurIPS 2025) |
| Replace the BADS weight network with individual scalar weights | CIFAR-10 · accuracy | Helped | A Bayesian Approach to Data Point Selection (NeurIPS 2024) |
What reviewers will expect
Datasets and baselines named in the 17 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 17 papers
- CIFAR-10 · 2 of 17 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 57 papers; from OpenAlex and the papers' affiliations).
- EPFL · 2 papers
- Texas A&M University - College Station · 2 papers
- RIKEN · 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
- Simulation-based inference (118 papers)
- Variational Bayesian deep learning (105 papers)
- Approximate inference in Bayesian neural networks (93 papers)
- Bayesian view of meta-learning (63 papers)
Elsewhere in the topic
- Priors and functional variational uncertainty (112 papers)
- Sparse variational Gaussian processes (92 papers)
- PAC-Bayes and decoupled GP learning (83 papers)
- Epistemic versus aleatoric decomposition (75 papers)
- Single-model predictive uncertainty (74 papers)
- Bayesian decision-making under uncertainty (57 papers)
- Infinite-width networks as Gaussian processes (49 papers)
- Regression uncertainty quantification (43 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).