Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Infinite-width networks as Gaussian processes
Connects wide neural networks and deep kernels to Gaussian processes.
49 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2017–2026) · an approach in Uncertainty and Gaussian processes, keyword cluster “gaussian processes · gp · kernel”
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.
10 papers in 2023–24 and 10 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 3.7% to 3.2%.
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
- 2017 · Structured Inference Networks for Nonlinear State Space Models (AAAI 2017 · 355 citations)
- 2018 · Deep Neural Networks as Gaussian Processes (ICLR 2018 · 553 citations)
- 2019 · Bayesian Deep Convolutional Networks with Many Channels are Gaussian Processes (ICLR 2019 · 149 citations)
- 2020 · Bayesian Deep Ensembles via the Neural Tangent Kernel (NeurIPS 2020 · 29 citations)
- 2021 · The Hintons in your Neural Network: a Quantum Field Theory View of Deep Learning (ICML 2021 · 6 citations)
- 2022 · Transformer Neural Processes: Uncertainty-Aware Meta Learning Via Sequence Modeling (ICML 2022 · 12 citations)
- 2023 · Autoregressive Conditional Neural Processes (ICLR 2023 · 3 citations)
- 2024 · Gaussian Process Neural Additive Models (AAAI 2024 · 11 citations)
- 2025 · Rényi Neural Processes (ICML 2025 · 0 citations)
- 2026 · ProEval: Proactive Failure Discovery and Efficient Performance Estimation for Generative AI Evaluation (ICML 2026 · 0 citations)
Earliest papers here
- Structured Inference Networks for Nonlinear State Space Models (AAAI 2017 · 355 citations)
- Random Feature Expansions for Deep Gaussian Processes (ICML 2017 · 75 citations)
- Deep Neural Networks as Gaussian Processes (ICLR 2018 · 553 citations)
What has been tried
42 papers here report 370 measured changes (a component added, removed or swapped, compared with a baseline): helped 61%, no clear effect 8%, hurt 23%, mixed 8%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Use L=100 rather than L=10 for HMC fine-tuning | Synthetic BN (0.3), 100 train epochs · MAE (× 10^-3) | 5.300 -> 6.864 | Hurt | On Neural Networks as Infinite Tree-Structured Probabilistic Graphical Models (NeurIPS 2024 · 1 citation) |
| Increase the number of Monte Carlo samples during training | RBF context and target · log-likelihood | Helped | Rényi Neural Processes (ICML 2025 · 0 citations) | |
| vary the kernel band-limit | synthetic regression · test L2 loss | Mixed | Infinite Neural Operators: Gaussian processes on functions (NeurIPS 2025 · 0 citations) | |
| Use Hodge GVFs with interdomain inducing variables instead of other tested synthetic-model variants | Synthetic sphere regression · NLPD | Helped | Residual Deep Gaussian Processes on Manifolds (ICLR 2025 · 0 citations) | |
| Increase depth from L=2 to L=3 | Simulation scenarios · RMSE | 0.57 (0.05) -> 0.56 (0.05); 0.86 (0.09) -> 0.82 (0.07); 8.08 (0.38) -> 8.09 (0.38) | No clear effect | Deep Kernel Posterior Learning under Infinite Variance Prior Weights (ICLR 2025 · 0 citations) |
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 20 papers; from OpenAlex and the papers' affiliations).
- University of Cambridge · 3 papers
- Capital One · 2 papers
Use it
Download BibTeX of 34 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
- Sparse variational Gaussian processes (92 papers)
- PAC-Bayes and decoupled GP learning (83 papers)
- Structured Gaussian process models (39 papers)
Elsewhere in the topic
- Nonparametric and PAC-Bayes Bayesian learning (130 papers)
- Simulation-based inference (118 papers)
- Priors and functional variational uncertainty (112 papers)
- Variational Bayesian deep learning (105 papers)
- Approximate inference in Bayesian neural networks (93 papers)
- Epistemic versus aleatoric decomposition (75 papers)
- Single-model predictive uncertainty (74 papers)
- Bayesian view of meta-learning (63 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).