Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Priors and functional variational uncertainty
Builds expressive priors, depth-based and functional variational methods for uncertainty estimation.
112 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2026) · an approach in Uncertainty and Gaussian processes, keyword cluster “uncertainty estimation · epistemic · quantification”
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.
35 papers in 2023–24 and 41 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic held steady, from 12.9% to 13.3%.
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 · DISCO Nets : DISsimilarity COefficients Networks (NeurIPS 2016 · 30 citations)
- 2018 · Dirichlet-based Gaussian Processes for Large-scale Calibrated Classification (NeurIPS 2018 · 33 citations)
- 2019 · Accurate Uncertainty Estimation and Decomposition in Ensemble Learning (NeurIPS 2019 · 31 citations)
- 2020 · LUVLi Face Alignment: Estimating Landmarks' Location, Uncertainty, and Visibility Likelihood (CVPR 2020 · 151 citations)
- 2021 · Uncertainty in Gradient Boosting via Ensembles (ICLR 2021 · 38 citations)
- 2022 · Deep Depth From Focus With Differential Focus Volume (CVPR 2022 · 43 citations)
- 2023 · Learning the Distribution of Errors in Stereo Matching for Joint Disparity and Uncertainty Estimation (CVPR 2023 · 48 citations)
- 2024 · Quantifying Uncertainty in Motion Prediction with Variational Bayesian Mixture (CVPR 2024 · 8 citations)
- 2025 · Auto Encoding Neural Process for Multi-interest Recommendation (AAAI 2025 · 2 citations)
- 2026 · Physics-Informed Multi-Task Learning for Battery State of Health Prediction with Uncertainty Quantification (AAAI 2026 · 2 citations)
Earliest papers here
- DISCO Nets : DISsimilarity COefficients Networks (NeurIPS 2016 · 30 citations)
- Structured Regression Gradient Boosting (CVPR 2016 · 4 citations)
- Dirichlet-based Gaussian Processes for Large-scale Calibrated Classification (NeurIPS 2018 · 33 citations)
What has been tried
79 papers here report 750 measured changes (a component added, removed or swapped, compared with a baseline): helped 68%, no clear effect 10%, hurt 18%, mixed 4%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| use 6 instead of 4 mixture components | INTERACTION val · minFDE6 | 0.5362 -> 0.4306 | Helped | Quantifying Uncertainty in Motion Prediction with Variational Bayesian Mixture (CVPR 2024 · 8 citations) |
| Use the reparameterization-based Hessian estimator instead of squared-gradient estimation | CIFAR-10, ResNet-20 · accuracy | 88.81 ± 0.31 -> 92.64 ± 0.13 | Helped | Variational Learning is Effective for Large Deep Networks (ICML 2024 · 3 citations) |
| Add the structure kernel | ProteinGym substitution, Avg. (174 assays) · Spearman | -0.065 | Helped | Kermut: Composite kernel regression for protein variant effects (NeurIPS 2024 · 3 citations) |
| add the neural process model | ML-100K · HR@10 | 15.49 -> 22.75 | Helped | Auto Encoding Neural Process for Multi-interest Recommendation (AAAI 2025 · 2 citations) |
| remove bi-Lipschitz regularization | CIFAR-10 ID vs CIFAR-100 OOD classification · ECE | 0.011±0.002 -> 0.042±0.001 | Hurt | Distance-informed Neural Processes (NeurIPS 2025 · 1 citation) |
| Add gradient-based shape refinement during Bayesian optimization | Airfoil optimization · optimized lift-to-drag ratio | Helped | Enabling Uncertainty Estimation in Iterative Neural Networks (ICML 2024 · 1 citation) | |
| remove 1-Lipschitz regularization | CIFAR-10-C defocus_blur_5 · accuracy | 91.2% -> 78.6% | Hurt | Density-Softmax: Efficient Test-time Model for Uncertainty Estimation and Robustness under Distribution Shifts (ICML 2024 · 1 citation) |
| Select inner points using K-means centroids rather than random sampling | FashionMNIST vs MNIST, 500 examples/class · AUROC | 0.9368 -> 0.9543 | Helped | Combining Statistical Depth and Fermat Distance for Uncertainty Quantification (NeurIPS 2024 · 1 citation) |
| Model accuracy and BART scores jointly with PTF rather than separately with PMF | Performance matrix, test ratio 20%, overall · RMSE | 0.175 -> 0.205 | Hurt | Can We Predict Performance of Large Models across Vision-Language Tasks? (ICML 2025 · 0 citations) |
| Use full-rank isotropic Gaussian noise instead of low-rank isotropic noise | WG-S · accuracy (%) | 77.72 -> 75.57 | Hurt | Training-Free Bayesianization for Low-Rank Adapters of Large Language Models (NeurIPS 2025 · 0 citations) |
| Increase PIA J from 10 to 100 | CIFAR100 vs SVHN · AUROC | 78.81± 1.87 -> 80.55± 1.99 | Helped | Credal Wrapper of Model Averaging for Uncertainty Estimation in Classification (ICLR 2025 · 0 citations) |
| Neglect non-existing edges in the modeled data-generating process | CSBM synthetic graphs · AUC | Hurt | Uncertainty for Active Learning on Graphs (ICML 2024 · 0 citations) |
What reviewers will expect
Datasets and baselines named in the 18 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 18 papers
- CIFAR-100 · 2 of 18 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 76 papers; from OpenAlex and the papers' affiliations).
- New York University · 4 papers
- Technical University Munich · 3 papers
- University of Oxford · 2 papers
- Peking University · 2 papers
- Korea Advanced Institute of Science and Technology · 2 papers
- Technische Universität München · 2 papers
- Stanford University · 2 papers
- NVIDIA · 2 papers
- Imperial College London · 2 papers
- IBM · 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
- Single-model predictive uncertainty (74 papers)
- Epistemic versus aleatoric decomposition (75 papers)
- Bayesian decision-making under uncertainty (57 papers)
Elsewhere in the topic
- Nonparametric and PAC-Bayes Bayesian learning (130 papers)
- Simulation-based inference (118 papers)
- Variational Bayesian deep learning (105 papers)
- Approximate inference in Bayesian neural networks (93 papers)
- Sparse variational Gaussian processes (92 papers)
- PAC-Bayes and decoupled GP learning (83 papers)
- Bayesian view of meta-learning (63 papers)
- Infinite-width networks as Gaussian processes (49 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).