Pith. sign in

REVIEW 5 major objections 6 minor 29 references

Evaluating Uncertainty in Deep Gaussian Processes

T0 review · 5 major / 6 minor · reviewed 2026-08-16 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Deep sigma point processes calibrate best in-distribution, but deep ensembles hold up better under distribution shift.

desk verdict First DSPP/ensemble calibration benchmark, but the DSPP scoring is underspecified and the headline numbers need tightening before I'd fully trust them. read the letter →

arxiv 2504.17719 v1 pith:FTTT2SNJ submitted 2025-04-24 stat.ML cs.LG

Matthijs van der Lende , Jeremias Lino Ferrao , Niclas Müller-Hof This is my paper · ORCID
classification stat.MLcs.LG
keywords deepGaussianprocessessigmapointuncertaintycalibrationexpectederrordistributionshiftensemblesnegativeloglikelihood
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper asks whether hierarchical Gaussian process models deliver what they promise in practice: well-calibrated uncertainty. It compares Deep Gaussian Processes (DGPs) and Deep Sigma Point Processes (DSPPs) against a Deep Ensemble of neural networks on a protein regression task and an EEG seizure classification task. The central finding is a split verdict: DSPPs achieve the lowest in-distribution calibration error on both datasets, while the Deep Ensemble holds its calibration and accuracy steady under synthetic feature perturbations that degrade the GP-based models. The result matters because it separates two goals—calibrated uncertainty on familiar data and safety under distribution shift—and shows they can point to different models.

What carries the argument

The central object is the Deep Sigma Point Process (DSPP), whose layers propagate uncertainty through learnable quadrature points (sigma points) instead of Monte Carlo sampling; this produces a finite Gaussian mixture predictive distribution (Eq 12) and enables direct maximum-likelihood training. The evaluation machinery consists of NLL and ECE metrics plus five synthetic feature-level perturbations (Gaussian noise, feature masking, feature scaling, feature permutation, outlier injection) applied at increasing severity.

What would settle it

Recompute DSPP NLL and ECE from the full Gaussian mixture of its predictive distribution instead of a collapsed single Gaussian; if Table 2 changes materially, the claimed calibration advantage is an artifact of the collapse. Independently, run the same models under real covariate shift datasets to see whether the ensemble's stability persists beyond the paper's five synthetic perturbations.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that DSPPs' sigma-point approximation yields excellent in-distribution calibration (ECE 0.026 on CASP, 0.035 on ESR) but that this advantage does not carry over to robustness: under the five feature-level shifts, the Deep Ensemble maintains roughly stable ECE and accuracy, whereas DGP and DSPP degrade—DGP sharply on regression. The paper concludes that good in-distribution calibration does not guarantee robustness, and that deep ensembles remain a strong baseline for deployment under shift.

Load-bearing premise

The paper evaluates DSPPs with a single mean and variance while its equations define a mixture predictive distribution; that unstated collapse is load-bearing for every calibration number reported.

Editorial extensions

If this is right

  • On tabular regression and classification, a DSPP is the better choice when in-distribution calibration is the priority.
  • For deployment where input features may be corrupted, a deep ensemble is the safer default because it held ECE near 0.11 on CASP and accuracy near 0.97 on ESR while the GP models degraded.
  • NLL alone is not enough to judge uncertainty quality; the severity curves of ECE reveal robustness differences that a single NLL average hides.
  • DSPP's sigma-point training works with a softmax likelihood, so its calibration benefit is not limited to Gaussian regression.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the goal is deployment under shift, an immediate next test is whether the ranking holds under label-preserving covariate shifts that change feature marginal distributions in ways the five synthetic perturbations do not.
  • The paper leaves implicit how the DSPP mixture predictive is reduced to the single mean and variance used in NLL and interval metrics; spelling out that reduction, or evaluating the mixture directly, would let others reproduce the calibration numbers.
  • The ensemble's stability suggests diversity across members, rather than Bayesian posterior propagation, protects calibration under feature corruption; comparing DSPPs with more sigma points would test whether the gap is architectural or inferential.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

5 major / 6 minor

Summary. This paper empirically compares Deep Gaussian Processes (DGPs), Deep Sigma Point Processes (DSPPs), and Deep Ensembles on two tabular benchmarks: regression on the CASP protein structure dataset and binary classification on the ESR seizure recognition dataset. It reports MAE/accuracy, NLL, and ECE for in-distribution performance, and evaluates ECE/MAE/accuracy under five synthetic feature-level perturbations at increasing severity. The authors conclude that DSPPs provide the strongest in-distribution calibration (ECE 0.026 on CASP, 0.035 on ESR), while Deep Ensembles are the most robust under shift, maintaining stable ECE and accuracy while GP-based methods degrade. The paper includes ablations on inducing points and depth, and makes code publicly available.

Significance. If the central results survive verification, this is a useful benchmark study: it extends the evaluation of DSPPs beyond NLL, applies DSPPs to classification with a softmax likelihood, and applies a uniform shift protocol across regression and classification. The provision of code, public datasets, and ablations strengthens reproducibility. However, the paper's quantitative claims are currently undermined by an unstated transformation of the mixture predictive distributions into the single-Gaussian NLL/interval formulas, by a likely misspecified ECE formula, and by the absence of error bars or statistical tests for the headline comparisons. These issues affect the central claims and require correction before the results can be interpreted.

major comments (5)
  1. [§2.3, §3.1, Eqs. (12), (19), (22)] The DSPP predictive is defined in Eq. (12) as a Q-component Gaussian mixture, and the DGP predictive is a Monte Carlo mixture, but the regression metrics in §3.1 are written for a single Gaussian: Eq. (19) evaluates NLL with one μ(x) and σ²(x), and Eq. (22) builds prediction intervals from the same two moments. The paper never states whether μ and σ² are the moments of the mixture, whether the exact mixture density is used for NLL, or whether interval calibration uses mixture quantiles. This is load-bearing because NLL is a proper scoring rule only for the exact predictive distribution, and interval-based ECE depends on the full CDF; if the mixture is collapsed to moments, the headline DSPP calibration numbers in Table 2 score a Gaussian approximation rather than the trained model, and if it is not, Eqs. (19) and (22) do not describe the evaluation. Please specify the exact evaluation procedure and recompute the affected numbers if needed.
  2. [§3.1, Eq. (21)] The ECE formula in Eq. (21) uses |B_b|/B as the bin weight instead of |B_b|/N, even though the bins partition the N test points and Σ_b |B_b| = N. With this formula the reported ECE values are not on the standard probability scale and are not comparable to the usual ECE definition; please correct the denominator and recompute Table 2 and Figures 4/7 if the implementation follows Eq. (21).
  3. [§3.5, Table 2, Figures 4 and 7] The paper does not report any measure of variability for Table 2, despite §3.5 stating that N = 5 independent training runs were performed per model for the shift experiments; it is unclear whether the Table 2 entries are single-seed results, averages, or best runs. The shift results are presented only as boxplots aggregating over runs and corruption types, with no confidence intervals, paired tests, or effect-size measures. As a consequence, the headline contrasts (DSPP ECE 0.026 vs. Deep Ensemble 0.112 on CASP; ensemble stability near 0.11 ECE under shift vs. GP degradation) cannot be distinguished from run-to-run variation. Please report means and standard deviations or CIs over the runs and add a statistical comparison (e.g., paired by seed and corruption type) for the shift outcomes.
  4. [§3.5, Eqs. (24)–(28), Figures 4 and 7] The shift analysis pools all five corruption types into a single boxplot per severity level, but the corruptions have qualitatively different effects—feature permutation preserves the marginal distribution of each column, whereas Gaussian noise, scaling, masking, and outliers change marginal or joint structure. Pooling conceals per-corruption differences that are needed to support the conclusion that Deep Ensembles are 'more robust' and GP methods 'vulnerable' in general. Please report per-corruption results (or include corruption type as a factor in the analysis) before drawing general robustness conclusions.
  5. [Appendix A, Table 3] The optimized DSPP for CASP has architecture [ ] (no hidden layers), so the headline in-distribution calibration advantage on the regression task (ECE 0.026) is achieved by a shallow sparse GP, not by a deep sigma point process. This limits the strength of the claim that 'DSPPs provide strong in-distribution calibration' for deep models on regression; please either include at least one deep DSPP configuration in the main CASP comparison or explicitly temper the depth-related interpretation.
minor comments (6)
  1. [§3.1, reference [27]] The citation [27] for the statement that NLL can overemphasize tail probabilities is Watkins and Dayan's Q-learning paper; the intended reference is likely [21]. Please correct.
  2. [§2.3] Typo: 'agumented' should be 'augmented'.
  3. [§5.1] Typo: 'This work focused on standard on the standard regression and classification' should read 'standard regression and classification'.
  4. [Figure 3 caption] Typo: 'Sezure' should be 'Seizure'.
  5. [Appendix C.2] The opening sentence, 'For this experiment, we were interested in the relationship between the number of and the negative log likelihood for our dataset,' is missing a word (likely 'layers'). Please fix.
  6. [Appendix B, Eq. (30)] The first displayed line of Eq. (30) is missing the logarithm on q_ψ(z); it should read E_{q_ψ(z)}[log q_ψ(z) − log(p_θ(x|z)p_θ(z)/p_θ(x))].

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: this is an empirical benchmarking study whose reported metrics are computed on held-out test data, not derived from fitted parameters or self-citations.

full rationale

The paper is a measurement study, not a derivation chain. The central claims are based on test-set NLL, ECE, MAE, and accuracy values reported in Table 2 and boxplots, computed against external baselines (Deep Ensembles) and public datasets (CASP and ESR). No quantity labeled a prediction is defined in terms of a fitted parameter, and no claimed result is equivalent by construction to an input. The DSPP predictive distribution is defined in Eq. (12) as a Gaussian mixture, while the metric formulas in Eqs. (19) and (22) are written for a single Gaussian; this creates an ambiguity about how the mixture is scored, but that is a reproducibility/correctness concern, not circularity, because the reported ECE and NLL are empirical evaluations rather than derivations that reduce to their inputs. The only self-citation is reference [17] in the Future Work section, which is used to mention prior work by an author on DGPs in reinforcement learning and is not load-bearing for any result in this paper. The DSPP model itself is cited to external work [14], and the conclusion that DSPPs are well calibrated in-distribution is independently supported by the held-out measurements reported here. Under the stated rules, a self-contained empirical evaluation against external benchmarks warrants a score of 0-2; no circular step meeting the required standard of exhibiting a specific reduction was found.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

No new theoretical constructs are introduced. The load-bearing inputs are the tuned hyperparameters and the unstated reduction of the DSPP mixture to Gaussian moments for metric computation. The synthetic shift scheme is an externally imposed stress test rather than part of the model.

free parameters (4)
  • KL divergence scaling beta = 1
    Fixed constant for DGP and DSPP objectives (Section 2.3); affects regularization strength but is not tuned.
  • DGP Monte Carlo samples = 10
    Fixed (Table 1); number of samples for the doubly stochastic ELBO approximation.
  • DSPP quadrature sites Q = 8
    Fixed (Table 1); number of sigma points in the Gaussian mixture estimator.
  • BayesOpt selected hyperparameters (learning rate, architecture, inducing points, ensemble size) = See Table 3 per model/dataset; e.g. DGP CASP: lr=0.1, M=159; DSPP CASP: lr=0.055, M=50
    Optimized by BayesOpt per model and dataset (Appendix A); the headline comparison is contingent on these model selection choices.
assumptions (4)
  • standard math Standard GP/SVGP/DGP/DSPP formulations are correct as presented.
    The paper relies on established GP theory from [6, 12, 14, 24] without re-deriving it.
  • domain assumption The five synthetic feature shifts in Section 3.5 are meaningful proxies for real-world distribution shift.
    The paper assumes Gaussian noise, masking, scaling, permutation, and outlier injection simulate realistic covariate shift; no real-world shifted dataset is used.
  • domain assumption The tabular datasets (CASP, ESR) are representative enough to support general claims about GP robustness.
    The paper notes limited generalizability to high-dimensional data in Section 5, but the abstract and conclusion still draw general conclusions about GP methods versus ensembles.
  • ad hoc to paper DSPP predictive distribution is reduced to a single Gaussian for NLL and ECE computation.
    Eq (12) defines a mixture but Eqs (19) and (22) use a single mean and variance; the paper never states how the mixture is collapsed, so the evaluation implicitly assumes this reduction.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Evaluating Uncertainty in Deep Gaussian Processes." pith.science (2026). https://pith.science/paper/FTTT2SNJ

@misc{pith2026250417719,
  author       = {Pith},
  title        = {Pith review of: Evaluating Uncertainty in Deep Gaussian Processes},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FTTT2SNJ}},
  note         = {Machine review of arXiv:2504.17719}
}
read the original abstract

Reliable uncertainty estimates are crucial in modern machine learning. Deep Gaussian Processes (DGPs) and Deep Sigma Point Processes (DSPPs) extend GPs hierarchically, offering promising methods for uncertainty quantification grounded in Bayesian principles. However, their empirical calibration and robustness under distribution shift relative to baselines like Deep Ensembles remain understudied. This work evaluates these models on regression (CASP dataset) and classification (ESR dataset) tasks, assessing predictive performance (MAE, Accu- racy), calibration using Negative Log-Likelihood (NLL) and Expected Calibration Error (ECE), alongside robustness under various synthetic feature-level distribution shifts. Results indicate DSPPs provide strong in-distribution calibration leveraging their sigma point approximations. However, compared to Deep Ensembles, which demonstrated superior robustness in both per- formance and calibration under the tested shifts, the GP-based methods showed vulnerabilities, exhibiting particular sensitivity in the observed metrics. Our findings underscore ensembles as a robust baseline, suggesting that while deep GP methods offer good in-distribution calibration, their practical robustness under distribution shift requires careful evaluation. To facilitate reproducibility, we make our code available at https://github.com/matthjs/xai-gp.

Figures

Figures reproduced from arXiv: 2504.17719 by the authors.

Figure 1
Figure 1. Training and validation loss curves of our optimized models on CASP and ESR [PITH_FULL_IMAGE:figures/full_fig_p010_1.png] view at source ↗
Figure 2
Figure 2. Calibration curves of our models for the Protein regression dataset. [PITH_FULL_IMAGE:figures/full_fig_p010_2.png] view at source ↗
Figure 3
Figure 3. Calibration curves of our models for the Epileptic Sezure Recognition dataset. [PITH_FULL_IMAGE:figures/full_fig_p011_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Box plots showing ECE under distributional shift for the CASP regression (left) and ESR [PITH_FULL_IMAGE:figures/full_fig_p011_4.png]
Figure 5
Figure 5. Figure 5: Negative log likelihood on test set against the number of inducing points. Left: CASP, [PITH_FULL_IMAGE:figures/full_fig_p017_5.png]
Figure 6
Figure 6. Figure 6: Negative log likelihood on test set against the number of hidden layers. Left: CASP, [PITH_FULL_IMAGE:figures/full_fig_p017_6.png]
Figure 7
Figure 7. Figure 7: Box plots showing MAE and accuracy for the ESR classification (right) and CASP [PITH_FULL_IMAGE:figures/full_fig_p018_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

29 extracted references · 17 canonical work pages

  1. [1]

    Indications of nonlinear deterministic and finite-dimensional structures in time series of brain electrical activity: Dependence on recording region and brain state

    Ralph Andrzejak, Klaus Lehnertz, Florian Mormann, Christoph Rieke, Peter David, and Christian Elger. Indications of nonlinear deterministic and finite-dimensional structures in time series of brain electrical activity: Dependence on recording region and brain state. Physical review. E, Statistical, nonlinear, and soft matter physics, 64:061907, 01 2002. d...

    work page doi:10.1103/physreve.64.061907 2002

  2. [2]

    Understanding probabilistic sparse gaussian process approximations.Advances in neural information processing systems, 29, 2016

    Matthias Bauer, Mark Van der Wilk, and Carl Edward Rasmussen. Understanding probabilistic sparse gaussian process approximations.Advances in neural information processing systems, 29, 2016. URL https://arxiv.org/abs/1606.04820

    arXiv 2016

  3. [3]

    Safe exploration in reinforcement learning: Theory and applications in robotics

    Felix Berkenkamp. Safe exploration in reinforcement learning: Theory and applications in robotics. PhD thesis, ETH Zurich, 2019. URL https://www.research-collection.ethz. ch/bitstream/handle/20.500.11850/370833/root.pdf

    work page 2019

  4. [4]

    Off-policy reinforcement learning with gaussian processes

    Girish Chowdhary, Miao Liu, Robert Grande, Thomas Walsh, Jonathan How, and Lawrence Carin. Off-policy reinforcement learning with gaussian processes. IEEE/CAA Journal of Automatica Sinica, 1(3):227–238, 2014. doi: 10.1109/JAS.2014.7004680

    work page arXiv 2014

  5. [5]

    D. R. Cox. Some statistical methods connected with series of events.Journal of the Royal Statistical Society: Series B (Methodological), 17(2):129–157, 12 2018. ISSN 0035-9246. doi: 10.1111/j.2517-6161.1955.tb00188.x. URL https://doi.org/10.1111/j.2517-6161.1955. tb00188.x

    arXiv 2018

  6. [6]

    Lawrence

    Andreas Damianou and Neil D. Lawrence. Deep Gaussian processes. In Carlos M. Carvalho and Pradeep Ravikumar, editors,Proceedings of the Sixteenth International Conference on Artificial Intelligence and Statistics, volume 31 ofProceedings of Machine Learning Research, pages 207–215, Scottsdale, Arizona, USA, 29 Apr–01 May 2013. PMLR. URL https:// proceedin...

    work page 2013

  7. [7]

    Reinforcement learning with gaussian processes

    Yaakov Engel, Shie Mannor, and Ron Meir. Reinforcement learning with gaussian processes. In Proceedings of the 22nd International Conference on Machine Learning, ICML ’05, page 201–208, New York, NY, USA, 2005. Association for Computing Machinery. ISBN 1595931805. doi: 10.1145/1102351.1102377. URL https://doi.org/10.1145/1102351.1102377

    arXiv 2005

  8. [8]

    Generalisation in humans and deep neural networks.Advances in neural information processing systems, 31, 2018

    Robert Geirhos, Carlos RM Temme, Jonas Rauber, Heiko H Schütt, Matthias Bethge, and Felix A Wichmann. Generalisation in humans and deep neural networks.Advances in neural information processing systems, 31, 2018

    2018

Show all 29 references
  1. [9]

    Sample efficient reinforcement learn- ing with gaussian processes

    Robert Grande, Thomas Walsh, and Jonathan How. Sample efficient reinforcement learn- ing with gaussian processes. In Eric P. Xing and Tony Jebara, editors, Proceedings of the 31st International Conference on Machine Learning, volume 32 ofProceedings of Ma- chine Learning Resea...

    2014

  2. [10]

    Benchmarking neural network robustness to common corruptions and perturbations.arXiv preprint arXiv:1903.12261, 2019

    Dan Hendrycks and Thomas Dietterich. Benchmarking neural network robustness to common corruptions and perturbations.arXiv preprint arXiv:1903.12261, 2019. 13

    1903 arXiv

  3. [11]

    Augmix: A simple data processing method to improve robustness and uncertainty

    Dan Hendrycks, Norman Mu, Ekin D Cubuk, Barret Zoph, Justin Gilmer, and Balaji Lakshmi- narayanan. Augmix: A simple data processing method to improve robustness and uncertainty. arXiv preprint arXiv:1912.02781, 2019

    1912 arXiv

  4. [12]

    Scalable Variational Gaussian Process Classification

    James Hensman, Alexander Matthews, and Zoubin Ghahramani. Scalable Variational Gaussian Process Classification. In Guy Lebanon and S. V. N. Vishwanathan, editors,Proceedings of the Eighteenth International Conference on Artificial Intelligence and Statistics, volume 38 of Proc...

    2015

  5. [13]

    Adversarial examples are not bugs, they are features.Advances in neural information processing systems, 32, 2019

    Andrew Ilyas, Shibani Santurkar, Dimitris Tsipras, Logan Engstrom, Brandon Tran, and Aleksander Madry. Adversarial examples are not bugs, they are features.Advances in neural information processing systems, 32, 2019

    2019

  6. [14]

    Deepsigmapointprocesses

    MartinJankowiak, GeoffPleiss, andJacobGardner. Deepsigmapointprocesses. InJonasPeters and David Sontag, editors,Proceedings of the 36th Conference on Uncertainty in Artificial Intelligence (UAI), volume 124 ofProceedings of Machine Learning Research, pages 789–798. PMLR, 03–06...

    2020

  7. [15]

    Reinforcement learning with gaussian process regression using variational free energy

    Kiseki Kameda and Fuyuhiko Tanaka. Reinforcement learning with gaussian process regression using variational free energy. Journal of Intelligent Systems, 32(1):20220205, 2023. doi: doi:10.1515/jisys-2022-0205. URL https://doi.org/10.1515/jisys-2022-0205

    2023 doi

  8. [16]

    Simple and scal- able predictive uncertainty estimation using deep ensembles

    Balaji Lakshminarayanan, Alexander Pritzel, and Charles Blundell. Simple and scal- able predictive uncertainty estimation using deep ensembles. In I. Guyon, U. Von Luxburg, S. Bengio, H. Wallach, R. Fergus, S. Vishwanathan, and R. Garnett, ed- itors, Advances in Neural Informa...

    2017

  9. [17]

    Interpretable function approximation with gaussian processes in value-based model-free reinforcement learning

    Matthijs van der Lende, Matthia Sabatelli, and Juan Cardenas-Cartagena. Interpretable function approximation with gaussian processes in value-based model-free reinforcement learning. In Tetiana Lutchyn, Adín Ramírez Rivera, and Benjamin Ricaud, editors,Pro- ceedings of the 6th...

    2025

  10. [18]

    A review of uncertainty for deep reinforcement learn- ing

    Owen Lockwood and Mei Si. A review of uncertainty for deep reinforcement learn- ing. Proceedings of the AAAI Conference on Artificial Intelligence and Interactive Digi- tal Entertainment, 18(1):155–162, Oct. 2022. doi: 10.1609/aiide.v18i1.21959. URL https: //ojs.aaai.org/index...

    2022 doi

  11. [19]

    Murphy.Probabilistic Machine Learning: Advanced Topics

    Kevin P. Murphy.Probabilistic Machine Learning: Advanced Topics. MIT Press, 2023. URL http://probml.github.io/book2

    2023

  12. [20]

    Dillon, Balaji Lakshminarayanan, and Jasper Snoek

    Yaniv Ovadia, Emily Fertig, Jie Ren, Zachary Nado, D Sculley, Sebastian Nowozin, Joshua V. Dillon, Balaji Lakshminarayanan, and Jasper Snoek. Can you trust your model’s uncertainty? evaluating predictive uncertainty under dataset shift, 2019. URLhttps://arxiv.org/abs/ 1906.02530

    2019 arXiv

  13. [21]

    Evaluating predictive uncertainty challenge

    Joaquin Quinonero-Candela, Carl Edward Rasmussen, Fabian Sinz, Olivier Bousquet, and Bernhard Schölkopf. Evaluating predictive uncertainty challenge. In Machine Learning Challenges Workshop, pages 1–27. Springer, 2005

    2005

  14. [22]

    Physicochemical Properties of Protein Tertiary Structure

    Prashant Rana. Physicochemical Properties of Protein Tertiary Structure. UCI Machine Learning Repository, 2013. DOI: https://doi.org/10.24432/C5QW3H

    2013 doi

  15. [23]

    Williams

    Carl Edward Rasmussen and Christopher K. Williams. Gaussian Processes for Machine Learning. The MIT Press, Cambridge, 2004. ISBN 978-0-262-25683-4. OCLC: 1178958074

    2004

  16. [24]

    Doubly stochastic variational inference for deep gaussian processes

    Hugh Salimbeni and Marc Deisenroth. Doubly stochastic variational inference for deep gaussian processes. In I. Guyon, U. Von Luxburg, S. Bengio, H. Wallach, R. Fergus, S. Vishwanathan, and 14 R. Garnett, editors,Advances in Neural Information Processing Systems, volume 30. Cur...

    2017

  17. [25]

    Dropout: a simple way to prevent neural networks from overfitting.The journal of machine learning research, 15(1):1929–1958, 2014

    Nitish Srivastava, Geoffrey Hinton, Alex Krizhevsky, Ilya Sutskever, and Ruslan Salakhutdinov. Dropout: a simple way to prevent neural networks from overfitting.The journal of machine learning research, 15(1):1929–1958, 2014

    1929

  18. [26]

    Variational learning of inducing variables in sparse gaussian processes

    Michalis Titsias. Variational learning of inducing variables in sparse gaussian processes. In David van Dyk and Max Welling, editors,Proceedings of the Twelfth International Conference on Artificial Intelligence and Statistics, volume 5 ofProceedings of Machine Learning Resear...

  19. [27]

    Q-learning.Machine learning, 8:279–292, 1992

    Christopher JCH Watkins and Peter Dayan. Q-learning.Machine learning, 8:279–292, 1992

    1992

  20. [28]

    Andrew Gordon Wilson, Zhiting Hu, Ruslan Salakhutdinov, and Eric P. Xing. Deep kernel learning. In Arthur Gretton and Christian C. Robert, editors,Proceedings of the 19th In- ternational Conference on Artificial Intelligence and Statistics, volume 51 ofProceedings of Machine L...

    2016

  21. [2009]

    URL https://proceedings.mlr.press/v5/titsias09a.html

    PMLR. URL https://proceedings.mlr.press/v5/titsias09a.html

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.