Pith. sign in

REVIEW 3 major objections 6 minor 1 cited by

Uncertainty quantification of neural network models of evolving processes via Langevin sampling

T0 review · 3 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 A Langevin sampling hypernetwork gives neural-ODE models uncertainty that tracks data variability more closely than standard variational inference in the demonstrated cases, with a cheaper last-layer variant preserving most of the fidelity.

desk verdict A useful, mostly sound UQ method for neural ODE models, but the time-factorized likelihood is a load-bearing approximation that the paper never tests. read the letter →

arxiv 2504.14854 v2 pith:5LQ2IMMH submitted 2025-04-21 cs.LG stat.ML

Cosmin Safta , Reese E. Jones , Ravi G. Patel , Raelynn Wonnacot , Dan S. Bolintineanu , Craig M. Hamel , Sharlotte L.B. Kramer This is my paper · ORCID
classification cs.LGstat.ML
keywords uncertaintyquantificationneuralordinarydifferentialequationsLangevinsamplinghypernetworkvariationalinferenceBayesiannetworksstochasticmaterialmodeling
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

The paper proposes a way to put principled uncertainty on neural-network models of evolving processes: it trains a sampler, a neural network that acts as the drift of a Langevin stochastic differential equation, to generate plausible weight vectors for the data model, whose terminal samples approximate the Bayesian posterior. The data model is a neural ordinary differential equation (NODE) for hidden state evolution coupled with an observation network, so the framework applies to history-dependent processes like chemical reactions and material response. The central claim, demonstrated on chemical kinetics and viscoelastic composites, is that this Langevin-sampling hypernetwork produces predictive uncertainty that tracks the variability in the data more faithfully than standard black-box variational inference, and that a cheaper variant with only the last network layers stochastic preserves most of that fidelity. A sympathetic reader would care because it offers a fully differentiable, scalable alternative to MCMC for attaching calibrated uncertainty bands to dynamical surrogate models used in engineering.

What carries the argument

The load-bearing object is the score-based Langevin sampler as a hypernetwork. A neural network $\mathrm{NN}_w(w;\theta)$ replaces the unknown score of the posterior $\nabla_w \log \pi(w\mid D)$, and its drift field, together with Brownian diffusion, defines an SDE whose stationary distribution is the target posterior. The prior is a unit-covariance Ornstein-Uhlenbeck process, so the KL divergence between prior and posterior random processes has the closed form in Eq. (14): an expectation over Langevin samples of $\tfrac{1}{2}\|\mathrm{NN}_w - f_0\|^2$ integrated in pseudo-time. This converts Bayesian inference into a differentiable optimization over $\theta$ with no acceptance step, and it also makes possible a divided parameter space $w = \{w^{(d)}, w^{(s)}\}$ in which some weights are sampled while others are re-optimized point estimates, the Bayesian last layer variant.

What would settle it

Run Algorithm 1 on a synthetic time series generated from the Langevin dynamics model with known parameters and with observation noise drawn from an AR(1) process with positive autocorrelation at fixed marginal variance; compare the inferred posterior variance of $(W,b)$ to the known analytical posterior. If the variance shrinks below the true value as autocorrelation increases, the independence assumption in Eq. (18) is the limiting assumption.

Watch

Extended reading notes

Core claim

The paper asserts that approximate posterior inference for NODE-based data models can be carried out by a hypernetwork in which a score network $\mathrm{NN}_w(w;\theta)$ drives an SDE $dw = \mathrm{NN}_w\,d\tau + \sqrt{2}\,dB_\tau$; after integration to pseudo-time $\tau_f$, the terminal weights are treated as samples from the posterior over the data-model parameters. The score network and the data model are trained jointly by maximizing an ELBO whose KL term is the path KL between the posterior SDE and an Ornstein-Uhlenbeck prior SDE, reducing to an expectation of the squared drift difference. On a classical Langevin model with an analytical posterior, the inferred joint density of the two data-model parameters matches the known covariance structure, whereas a full-covariance BBVI surrogate overestimates parameter variances by roughly 70% and distorts their correlation. On the bistable Schl\"ogl reaction, the LS ensemble captures both modes of the response density while diagonal-covariance BBVI explores only the wider mode. On material data, restricting stochasticity to the last layers and re-optimizing the remaining parameters gives uncertainty comparable to a fully stochastic treatment at lower sampling cost.

Load-bearing premise

The load-bearing premise is that observations at different time steps can be treated as independent (Eq. 18); if the residuals of the true process are correlated in time, this product likelihood overstates evidence and the posterior uncertainty bands will be too narrow.

Editorial extensions

If this is right

  • In the Langevin-dynamics exemplar, the posterior over the two model parameters $(W,b)$ recovered by LS matches the analytical marginal and joint statistics, while BBVI overestimates the standard deviations and inflates the correlation.
  • On the Schl\"ogl reaction, the LS ensemble reproduces the bimodal response distribution, whereas the diagonal-covariance BBVI baseline collapses onto the wider mode.
  • On hyperelastic material data, the Bayesian last layer with re-optimized deterministic parameters achieves a Wasserstein-1 distance 0.71 times that of the full stochastic treatment, while fixing the deterministic parameters at their MLE values makes it 1.18 times.
  • The framework lets the user tune the stochastic/deterministic split of the network parameters, trading sampling cost against uncertainty fidelity, and the convergence plots show the result is fairly insensitive to the number of replicas and SDE steps once past a modest threshold.
  • Because the sampler is fully differentiable and has no acceptance criterion, it can be embedded in a host simulation or optimal experimental design loop.

Reading between the lines

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

  • Extending the likelihood to account for serially correlated residuals, which the paper flags in Eq. (18) as an approximation, would change the inferred posterior widths; a testable version is to feed AR(1) noise with known autocorrelation and check whether uncertainty bands widen appropriately.
  • The success of the last-layer-plus-reoptimization scheme suggests a general design rule for deep surrogate models: if the inner layers are free to adapt, only the output layer needs to carry the posterior, a hypothesis that could be checked on larger-scale NODE architectures.
  • The amortized nature of the learned score network points toward online settings, where new trajectory data could update the posterior by continuing to train $\theta$ rather than re-running MCMC.
  • The same score-SDE construction could incorporate physics constraints, such as weight bounds or convexity, by restricting the output of $\mathrm{NN}_w$; the paper lists this as future work.
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

3 major / 6 minor

Summary. The paper proposes a hypernetwork framework for Bayesian uncertainty quantification of neural-network models of history-dependent processes. The data model is a neural ODE for the hidden state evolution together with a trainable observation model. The posterior over a subset of model weights is represented by terminal samples of an SDE whose drift is a learned score network, and the training objective is an ELBO whose KL term is a pathwise KL between prior and posterior SDEs. The method is demonstrated on a synthetic Langevin-dynamics example with known analytical posterior statistics, on a Schlögl reaction network with multimodal response, and on a viscoelastic composite material model, with comparisons to black-box variational inference and Bayesian-last-layer variants.

Significance. If the central claims hold, the paper presents a scalable and fully differentiable alternative to MCMC and standard VI for UQ in NODE-based models. Strengths include a concrete algorithm (Alg. 1), a nontrivial external benchmark with known joint posterior statistics in Sec. 4.1, a pathwise KL derivation in App. A that is largely correct, and a Bayesian-last-layer extension that reduces sampling cost while retaining predictive bands comparable to the fully stochastic treatment in Sec. 4.2. The demonstrations show that the Langevin hypernetwork can capture multimodal response densities that diagonal-covariance BBVI misses. However, several formula-level inconsistencies in the objective and a strong temporal-independence assumption in the likelihood currently limit the strength of the claims about trajectory-level uncertainty.

major comments (3)
  1. [Sec. 3, Eq. (18); Sec. 4] The likelihood in Eq. (18) factorizes over time points within each trajectory. All demonstrations use data that are strongly autocorrelated in time: SDE trajectories in Sec. 4.1 and random-walk-driven viscoelastic responses in Sec. 4.2. For the Langevin-dynamics example, the OU autocorrelation between adjacent output times is exp(-gamma*Delta_t) ≈ 0.92, so residuals at neighboring times are far from independent. A product of per-time marginals overstates the evidence and systematically narrows the posterior and predictive bands. The paper concedes immediately after Eq. (18) that this approximation can be relaxed, but all reported evaluations (W1 distances, KDE densities, per-time standard-deviation bands) are time-marginal and cannot detect the resulting misspecification of joint or trajectory-level uncertainty. Because the central claim is that the method produces predictive uncertainty matching data variability for evolving processes, this approximation is load-bearing. Please either add an experiment or analysis with a multivariate likelihood that models temporal residual correlation, or explicitly reframe the claims as marginal-calibration claims and justify the independence likelihood for that purpose.
  2. [Sec. 3, Eqs. (10)-(12)] The drift of the posterior SDE is written as f* = grad_w pi(w|D) and called 'the score of the posterior.' The score is the gradient of the log density, grad_w log pi(w|D). Eq. (8) correctly uses the log posterior, and App. C derives derivatives of log densities. As written, Eqs. (10)-(12) are inconsistent with Eq. (9) and with the KL expression in Eq. (14), which are based on drifts of SDEs. Please correct Eqs. (10)-(12) and the surrounding text to grad_w log pi(w|D), and state explicitly that the network NN_w in Eq. (12) approximates this log-density gradient.
  3. [Sec. 3, Eq. (14); App. A, Eq. (A-16)] For the SDE dw = f dτ + gamma*sqrt(2) dBτ, the transition covariance in App. A is ΣΔτ with Σ = 2γ²I, so the pathwise KL in Eq. (A-16) contributes (1/2) E[||f* - f0||²_Σ] = E[||f* - f0||²/(4γ²)]. Eq. (14) gives E[||f* - f0||²/(2γ²)], a factor of two larger. Because the KL term balances the likelihood term in the ELBO, this changes the objective actually optimized. Please align Eq. (14) with App. A (for example, by replacing γ with sqrt(2)γ or by redefining the diffusion normalization in Eq. (3)). Additionally, Eq. (14) omits the initial-condition term D_KL(q0||p0) that appears in Eq. (A-16); if the initial distributions are fixed this term is constant with respect to θ, but the stated equality should be amended accordingly.
minor comments (6)
  1. [Sec. 3, after Eq. (9)] The text says 'Weiner process'; this should be 'Wiener process.'
  2. [App. A, around Eq. (A-1) and Eq. (A-13)] The notation 'ΔB_k ∝ N(0, sqrt(Δτ) I)' is imprecise; a Brownian increment over Δτ is normally distributed as ΔB_k ∼ N(0, Δτ I). Please use standard increment notation to avoid confusion.
  3. [Sec. 3, Eq. (21)] The observation noise variance σ² is estimated from the same data used for training (σ_D from the data spread and σ_MLE from the MLE fit). This is an empirical-Bayes-like procedure; please discuss the potential effect of using the data twice and, if feasible, include a sensitivity analysis of the reported W1 metrics to the noise estimate.
  4. [Sec. 4.2.1, Fig. 11 caption] The word 'converage' in the caption should be 'coverage.'
  5. [Sec. 4.2, Eq. (36)] The text says 'analogous to Eq. (36)' when referring to the evolution equation for the internal state; the intended reference is likely Eq. (4), not Eq. (36).
  6. [Sec. 5] The concluding sentence describes the method as 'a competing alternative to ensemble methods such as the Stein variational gradient descent algorithm,' but no numerical comparison to SVGD is provided in this paper; please temper this claim or add a supporting reference/experiment.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the ELBO-trained Langevin hypernetwork is self-contained, with an external analytical benchmark and only non-load-bearing self-citations.

full rationale

The paper's derivation is a variational approximation: the sampler q is defined by the SDE in Eq. (3) with a learned drift NNw(w;theta), and the optimization loss in Eq. (13) is the ELBO with the path-KL term in Eq. (14). No predicted quantity, such as posterior weight statistics, predictive densities, W1 distances, or parity bands, is set equal to a fitted parameter or to an input of the loss by construction. The analytical Langevin example in Sec. 4.1 provides an external benchmark with closed-form score and moments, so the comparison is not a self-fulfilling fit. The prior OU process, the score-network parameterization, and the path-KL bound are standard results cited to external literature (Refs. [10,17-20,72-74]). Citations to the authors' own prior work ([60,65,66]) are used only to supply the material-physics data model or to name alternative algorithms, not to justify the core inference claim. The likelihood factorization in Eq. (18) is a stated approximation with a flagged limitation, and the noise variances in Eq. (21) are estimated from data; these make the predictive bands consistent with the data scale by assumption rather than by a hidden redefinition of the output. These are correctness and modeling concerns, not circularity. The score of 1 reflects minor self-citations that are not load-bearing.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The method relies on a handful of hand-set hyperparameters (noise variances, mixture weights, diffusion coefficient) and distributional assumptions (independence across time, shared diffusion, terminal sample sufficiency). No new physical entities are introduced.

free parameters (5)
  • observation noise variance sigma^2 = estimated per time step and mode via Eq. (21)
    Used in the Gaussian mixture likelihood (19)-(20). Estimated from data scatter plus MLE residual, not inferred from the posterior.
  • mixture weights alpha_l = counted from data trajectories for Nmodes=2
    Mode weights in the likelihood are set by counting data, not learned.
  • prior OU mean wbar = zero or MLE
    Mean of the OU prior; choice affects shrinkage of posterior samples.
  • SDE diffusion coefficient gamma = 1, sometimes annealed from 2 to 1
    Diffusion magnitude in Eqs. (3) and (22); fixed by hand rather than inferred.
  • KL weighting factor = implicitly 1 in Eq. (14)
    If Eq. (14) has a normalization error relative to App. A, the KL penalty scale is effectively an unstated free choice.
assumptions (4)
  • domain assumption Prior and posterior SDEs share the same diffusion coefficient (gamma0=gamma*)
    Required for finite KL and Eq. (14). Stated in Sec. 3 before Eq. (14).
  • domain assumption Observations across time steps are independent (Eq. 18)
    Likelihood factorizes over time, ignoring temporal correlation in residuals.
  • standard math Euler-Maruyama discretization provides unbiased ELBO gradient estimates
    Assumed without proof; standard in SDE variational inference.
  • domain assumption Terminal samples w(tao_f) approximate posterior draws q(w;theta)
    Used in Eq. (15) to estimate the likelihood term; requires mixing within tao_f.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Uncertainty quantification of neural network models of evolving processes via Langevin sampling." pith.science (2026). https://pith.science/paper/5LQ2IMMH

@misc{pith2026250414854,
  author       = {Pith},
  title        = {Pith review of: Uncertainty quantification of neural network models of evolving processes via Langevin sampling},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5LQ2IMMH}},
  note         = {Machine review of arXiv:2504.14854}
}
read the original abstract

We propose a scalable, approximate inference hypernetwork framework for a general model of history-dependent processes. The flexible data model is based on a neural ordinary differential equation (NODE) representing the evolution of internal states together with a trainable observation model subcomponent. The posterior distribution corresponding to the data model parameters (weights and biases) follows a stochastic differential equation with a drift term related to the score of the posterior that is learned jointly with the data model parameters. This Langevin sampling approach offers flexibility in balancing the computational budget between the evaluation cost of the data model and the approximation of the posterior density of its parameters. We demonstrate performance of the ensemble sampling hypernetwork on chemical reaction and material physics data and compare it to standard variational inference.

Figures

Figures reproduced from arXiv: 2504.14854 by the authors.

Figure 1
Figure 1. Schematics: (a) hypernetwork consisting of subcomponents: (b) sampler [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Langevin dynamics: (a) data density, (b) predicted density via Langevin sampling (Algorithm 1), and (c) [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. Langevin model: kernel density estimates of 1D marginal distributions and 2D joint densities for model [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figures from the paper (11 more)
Figure 4
Figure 4. Figure 4: Langevin dynamics: Wasserstein W1 distance as function of time and epochs for sequence of samples. 0.0 0.2 0.4 0.6 0.8 1.0 TIME 0.0 0.5 1.0 1.5 2.0 2.5 3.0 CONCENTRATION (a) data trajectories 0.0 0.2 0.4 0.6 0.8 1.0 TIME 0.0 0.5 1.0 1.5 2.0 2.5 3.0 CONCENTRATION (b) LS…
Figure 5
Figure 5. Figure 5: Schlogl reaction: trajectories (upper panels) and density (lower panels). ¨ hovering around 0.1. We conjecture that this is due to the fungibility of network weights which can result in several processing pathways contributing a fraction of the output. 4.2 Material phy…
Figure 6
Figure 6. Figure 6: Schlogl reaction: Wasserstein ¨ W1 distance as function of number of replica samples for sequence of sample pseudo-time steps. 0 50 100 150 200 250 300 Weight ID 0 50 100 150 200 250 300 Weight ID 0.08 0.10 0.12 0.14 0.16 0.18 0.20 (a) Distance Correlation 0.08 0.12 0.…
Figure 7
Figure 7. Figure 7: Distance correlation values for the parameters of (a) [PITH_FULL_IMAGE:figures/full_fig_p011_7.png]
Figure 8
Figure 8. Figure 8: Viscoelastic matrix with elastic inclusions. Three realizations in biaxial stretch. Stress is indicated by color [PITH_FULL_IMAGE:figures/full_fig_p011_8.png]
Figure 9
Figure 9. Figure 9: Hyperelastic data: parity plot for a MLE model fit (a) training data, (b) validation data. Black circles show [PITH_FULL_IMAGE:figures/full_fig_p013_9.png]
Figure 10
Figure 10. Figure 10: Hyperelastic: SDE trajectories for select parameters of [PITH_FULL_IMAGE:figures/full_fig_p013_10.png]
Figure 11
Figure 11. Figure 11: Hyperelastic data: parity plot for a stochastic model fit of the validation data. The shaded area shows [PITH_FULL_IMAGE:figures/full_fig_p014_11.png]
Figure 12
Figure 12. Figure 12: Viscoelastic composite data: MLE fit of a NODE model through the viscoelastic data held out for validation. [PITH_FULL_IMAGE:figures/full_fig_p014_12.png]
Figure 13
Figure 13. Figure 13: Viscoelastic composite response results for selected held-out validation random walk trajectories. The [PITH_FULL_IMAGE:figures/full_fig_p015_13.png]
Figure 14
Figure 14. Figure 14: Distance correlation values for the parameters of (a) [PITH_FULL_IMAGE:figures/full_fig_p016_14.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Differentiable neural network representation of multi-well, locally-convex potentials

    stat.ML 2025-06 conditional novelty 6.0 of 10

    A log-sum-exp mixture of input-convex neural networks fits multi-well potentials with learnable transition sharpness and sparse mode discovery.

Reference graph

Works this paper leans on

75 extracted references · 41 canonical work pages · cited by 1 Pith paper

  1. [1]

    Observers for multivariable systems

    David Luenberger. Observers for multivariable systems. IEEE transactions on automatic control, 11(2):190–197, 1966. 16 UQ of NN models via Langevin sampling A PREPRINT

    work page 1966

  2. [2]

    Neural ordinary differential equa- tions

    Ricky TQ Chen, Yulia Rubanova, Jesse Bettencourt, and David K Duvenaud. Neural ordinary differential equa- tions. Advances in neural information processing systems, 31, 2018

    work page 2018

  3. [3]

    Augmented neural odes

    Emilien Dupont, Arnaud Doucet, and Yee Whye Teh. Augmented neural odes. Advances in neural information processing systems, 32, 2019

    2019

  4. [4]

    Practical Markov chain Monte Carlo

    Charles J Geyer. Practical Markov chain Monte Carlo. Statistical science, pages 473–483, 1992

    work page 1992

  5. [5]

    Markov chain Monte Carlo in practice

    Walter R Gilks, Sylvia Richardson, and David Spiegelhalter. Markov chain Monte Carlo in practice. CRC press, 1995

    1995

  6. [6]

    Markov chain monte carlo method and its application

    Stephen Brooks. Markov chain monte carlo method and its application. Journal of the royal statistical society: series D (the Statistician), 47(1):69–100, 1998

    work page 1998

  7. [7]

    Variational inference: A review for statisticians

    David M Blei, Alp Kucukelbir, and Jon D McAuliffe. Variational inference: A review for statisticians. Journal of the American statistical Association, 112(518):859–877, 2017

    2017

  8. [8]

    Practical variational inference for neural networks

    Alex Graves. Practical variational inference for neural networks. Advances in neural information processing systems, 24, 2011

    2011

Show all 75 references
  1. [9]

    Stein variational gradient descent: A general purpose Bayesian inference algorithm

    Qiang Liu and Dilin Wang. Stein variational gradient descent: A general purpose Bayesian inference algorithm. Advances in neural information processing systems, 29, 2016

    2016

  2. [10]

    Bayesian learning via stochastic gradient Langevin dynamics

    Max Welling and Yee W Teh. Bayesian learning via stochastic gradient Langevin dynamics. In Proceedings of the 28th international conference on machine learning (ICML-11), pages 681–688. Citeseer, 2011

    2011

  3. [11]

    Riemann manifold Langevin and Hamiltonian Monte Carlo methods

    Mark Girolami and Ben Calderhead. Riemann manifold Langevin and Hamiltonian Monte Carlo methods. Jour- nal of the Royal Statistical Society Series B: Statistical Methodology, 73(2):123–214, 2011

    2011

  4. [12]

    The geometric foundations of hamil- tonian monte carlo

    Michael Betancourt, Simon Byrne, Sam Livingstone, and Mark Girolami. The geometric foundations of hamil- tonian monte carlo. Bernoulli, pages 2257–2298, 2017

    2017

  5. [13]

    A conceptual introduction to Hamiltonian Monte Carlo

    Michael Betancourt. A conceptual introduction to Hamiltonian Monte Carlo. arXiv preprint arXiv:1701.02434, 2017

    2017 arXiv

  6. [14]

    Understanding molecular simulation: from algorithms to applications

    Daan Frenkel and Berend Smit. Understanding molecular simulation: from algorithms to applications. Elsevier, 2023

    2023

  7. [15]

    ODE2V AE: Deep generative second order ODEs with Bayesian neural networks

    Cagatay Yildiz, Markus Heinonen, and Harri Lahdesmaki. ODE2V AE: Deep generative second order ODEs with Bayesian neural networks. In H. Wallach, H. Larochelle, A. Beygelzimer, F. d'Alch´e-Buc, E. Fox, and R. Garnett, editors, Advances in Neural Information Processing Systems, ...

    2019

  8. [16]

    Bayesian neural ordinary differential equations

    Raj Dandekar, Karen Chung, Vaibhav Dixit, Mohamed Tarek, Aslan Garcia-Valadez, Krishna Vishal Vemula, and Chris Rackauckas. Bayesian neural ordinary differential equations. arXiv preprint arXiv:2012.07244, 2020

    2012 arXiv

  9. [17]

    Winnie Xu, Ricky T. Q. Chen, Xuechen Li, and David Duvenaud. Infinitely Deep Bayesian Neural Networks with Stochastic Differential Equations. In Gustau Camps-Valls, Francisco J. R. Ruiz, and Isabel Valera, edi- tors, Proceedings of The 25th International Conference on Artifici...

    2022

  10. [18]

    Variational inference for stochastic differential equations

    Manfred Opper. Variational inference for stochastic differential equations. Annalen der Physik, 531(3):1800233, 2019

    2019

  11. [19]

    Scalable gradients for stochastic differential equations

    Xuechen Li, Ting-Kam Leonard Wong, Ricky TQ Chen, and David Duvenaud. Scalable gradients for stochastic differential equations. In International Conference on Artificial Intelligence and Statistics , pages 3870–3882. PMLR, 2020

    2020

  12. [20]

    Neural stochastic differential equations: Deep latent gaussian models in the diffusion limit

    Belinda Tzen and Maxim Raginsky. Neural stochastic differential equations: Deep latent gaussian models in the diffusion limit. arXiv preprint arXiv:1905.09883, 2019

    1905 arXiv

  13. [21]

    Gaussian processes for regression

    Christopher Williams and Carl Rasmussen. Gaussian processes for regression. Advances in neural information processing systems, 8, 1995

    1995

  14. [22]

    Flow matching for generative modeling

    Yaron Lipman, Ricky TQ Chen, Heli Ben-Hamu, Maximilian Nickel, and Matt Le. Flow matching for generative modeling. arXiv preprint arXiv:2210.02747, 2022

    2022 arXiv

  15. [23]

    Score- based generative modeling through stochastic differential equations

    Yang Song, Jascha Sohl-Dickstein, Diederik P Kingma, Abhishek Kumar, Stefano Ermon, and Ben Poole. Score- based generative modeling through stochastic differential equations. arXiv preprint arXiv:2011.13456, 2020

    2011 arXiv

  16. [24]

    Maximum likelihood training of score-based diffusion models

    Yang Song, Conor Durkan, Iain Murray, and Stefano Ermon. Maximum likelihood training of score-based diffusion models. Advances in neural information processing systems, 34:1415–1428, 2021

    2021

  17. [25]

    Score-based diffusion models via stochastic differential equations–a technical tutorial

    Wenpin Tang and Hanyang Zhao. Score-based diffusion models via stochastic differential equations–a technical tutorial. arXiv preprint arXiv:2402.07487, 2024. 17 UQ of NN models via Langevin sampling A PREPRINT

    2024 arXiv

  18. [26]

    Stochastic interpolants: A unifying framework for flows and diffusions

    Michael S Albergo, Nicholas M Boffi, and Eric Vanden-Eijnden. Stochastic interpolants: A unifying framework for flows and diffusions. arXiv preprint arXiv:2303.08797, 2023

    2023 arXiv

  19. [27]

    Score Operator Newton transport

    Nisha Chandramoorthy, Florian T Schaefer, and Youssef M Marzouk. Score Operator Newton transport. In International Conference on Artificial Intelligence and Statistics, pages 3349–3357. PMLR, 2024

    2024

  20. [28]

    Learning and using scores for efficient Bayesian filtering

    Nisha Chandramoorthy, Youssef Marzouk, and Anant Gupta. Learning and using scores for efficient Bayesian filtering. Bulletin of the American Physical Society, 2024

    2024

  21. [29]

    Interacting Langevin diffusions: Gradient structure and ensemble Kalman sampler.SIAM Journal on Applied Dynamical Systems, 19(1):412–441, 2020

    Alfredo Garbuno-Inigo, Franca Hoffmann, Wuchen Li, and Andrew M Stuart. Interacting Langevin diffusions: Gradient structure and ensemble Kalman sampler.SIAM Journal on Applied Dynamical Systems, 19(1):412–441, 2020

    2020

  22. [30]

    Hypernetworks

    David Ha, Andrew Dai, and Quoc V Le. Hypernetworks. arXiv preprint arXiv:1609.09106, 2016

    2016 arXiv

  23. [31]

    A brief review of hypernetworks in deep learning

    Vinod Kumar Chauhan, Jiandong Zhou, Ping Lu, Soheila Molaei, and David A Clifton. A brief review of hypernetworks in deep learning. Artificial Intelligence Review, 57(9):250, 2024

    2024

  24. [32]

    Bayesian hypernetworks

    David Krueger, Chin-Wei Huang, Riashat Islam, Ryan Turner, Alexandre Lacoste, and Aaron Courville. Bayesian hypernetworks. arXiv preprint arXiv:1710.04759, 2017

    2017 arXiv

  25. [33]

    Hypernetworks in the science of complex systems, volume 3

    Jeffrey Johnson. Hypernetworks in the science of complex systems, volume 3. World Scientific, 2013

    2013

  26. [34]

    Hypernetworks in meta- reinforcement learning

    Jacob Beck, Matthew Thomas Jackson, Risto Vuorio, and Shimon Whiteson. Hypernetworks in meta- reinforcement learning. In Conference on Robot Learning, pages 1478–1487. PMLR, 2023

    2023

  27. [35]

    Evolving hypernetwork model.The European Physical Journal B, 77:493–498, 2010

    Jian-Wei Wang, Li-Li Rong, Qiu-Hong Deng, and Ji-Yong Zhang. Evolving hypernetwork model.The European Physical Journal B, 77:493–498, 2010

    2010

  28. [36]

    Graph hypernetworks for neural architecture search

    Chris Zhang, Mengye Ren, and Raquel Urtasun. Graph hypernetworks for neural architecture search. arXiv preprint arXiv:1810.05749, 2018

    2018 arXiv

  29. [37]

    Stochastic hyperparameter optimization through hypernetworks

    Jonathan Lorraine and David Duvenaud. Stochastic hyperparameter optimization through hypernetworks. arXiv preprint arXiv:1802.09419, 2018

    2018 arXiv

  30. [38]

    Meta-learning via hypernetworks

    Dominic Zhao, Seijin Kobayashi, Jo˜ao Sacramento, and Johannes von Oswald. Meta-learning via hypernetworks. In 4th Workshop on Meta-Learning at NeurIPS 2020 (MetaLearn 2020). NeurIPS, 2020

    2020

  31. [39]

    Second order derivatives for network pruning: Optimal brain surgeon.Advances in neural information processing systems, 5, 1992

    Babak Hassibi and David Stork. Second order derivatives for network pruning: Optimal brain surgeon.Advances in neural information processing systems, 5, 1992

    1992

  32. [40]

    Optimal brain surgeon and general network pruning

    Babak Hassibi, David G Stork, and Gregory J Wolff. Optimal brain surgeon and general network pruning. In IEEE international conference on neural networks, pages 293–299. IEEE, 1993

    1993

  33. [41]

    Adam: A method for stochastic optimization

    Diederik P Kingma and Jimmy Ba. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980, 2014

    2014 arXiv

  34. [42]

    Marginalized neural network mixtures for large-scale regression

    Miguel L ´azaro-Gredilla and An ´ıbal R Figueiras-Vidal. Marginalized neural network mixtures for large-scale regression. IEEE transactions on neural networks, 21(8):1345–1351, 2010

    2010

  35. [43]

    Latent derivative Bayesian last layer networks

    Joe Watson, Jihao Andreas Lin, Pascal Klink, Joni Pajarinen, and Jan Peters. Latent derivative Bayesian last layer networks. In International Conference on Artificial Intelligence and Statistics , pages 1198–1206. PMLR, 2021

    2021

  36. [44]

    Jordan, Zoubin Ghahramani, Tommi S

    Michael I. Jordan, Zoubin Ghahramani, Tommi S. Jaakkola, and Lawrence K. Saul. An introduction to variational methods for graphical models. Machine Learning, 37(2):183–233, 1999

    1999

  37. [45]

    Black Box Variational Inference

    Rajesh Ranganath, Sean Gerrish, and David Blei. Black Box Variational Inference. In Samuel Kaski and Jukka Corander, editors, Proceedings of the Seventeenth International Conference on Artificial Intelligence and Statis- tics, volume 33 of Proceedings of Machine Learning Resea...

  38. [46]

    An introduction to stochastic processes in physics

    Don S Lemons and Paul Langevin. An introduction to stochastic processes in physics. JHU Press, 2002

    2002

  39. [47]

    World Scientific, 2012

    William Coffey and Yu P Kalmykov.The Langevin equation: with applications to stochastic problems in physics, chemistry and electrical engineering, volume 27. World Scientific, 2012

    2012

  40. [48]

    Generalized Langevin Equations and Many-Body Problems in Chemical Dynamics

    SA Adelman. Generalized Langevin Equations and Many-Body Problems in Chemical Dynamics. Advances in Chemical Physics, pages 143–253, 1980

    1980

  41. [49]

    Thermostat algorithms for molecular dynamics simulations

    Philippe H H ¨unenberger. Thermostat algorithms for molecular dynamics simulations. Advanced computer sim- ulation: Approaches for soft matter sciences I, pages 105–149, 2005. 18 UQ of NN models via Langevin sampling A PREPRINT

    2005

  42. [50]

    Temperature control in molecular dynamic simulations of non-equilibrium processes

    Dawid Toton, Christian D Lorenz, Nikolaos Rompotis, Natalia Martsinovich, and Lev Kantorovich. Temperature control in molecular dynamic simulations of non-equilibrium processes. Journal of Physics: Condensed Matter, 22(7):074205, 2010

    2010

  43. [51]

    Generalized Langevin equation approach for atom/solid-surface scattering: Collinear atom/harmonic chain model

    SA Adelman and J D Doll. Generalized Langevin equation approach for atom/solid-surface scattering: Collinear atom/harmonic chain model. The Journal of Chemical Physics, 61(10):4242–4245, 1974

    1974

  44. [52]

    Generalized Langevin equation approach for atom/solid-surface scattering: general formulation for classical scattering off harmonic solids

    SA Adelman and JD Doll. Generalized Langevin equation approach for atom/solid-surface scattering: general formulation for classical scattering off harmonic solids. The Journal of chemical physics , 64(6):2375–2388, 1976

    1976

  45. [53]

    Coupling of atomistic and continuum simulations using a bridging scale decomposition

    Gregory J Wagner and Wing Kam Liu. Coupling of atomistic and continuum simulations using a bridging scale decomposition. Journal of Computational Physics, 190(1):249–274, 2003

    2003

  46. [54]

    On the statistical calibration of physical models

    Khachik Sargsyan, Habib N Najm, and Roger Ghanem. On the statistical calibration of physical models. Inter- national Journal of Chemical Kinetics, 47(4):246–276, 2015

    2015

  47. [55]

    Schl ¨ogl

    F. Schl ¨ogl. Chemical reaction models for non-equilibrium phase transitions. Zeitschrift f¨ur Physik, 253(2):147– 161, 1972

    1972

  48. [56]

    Stochastic dynamics and non-equilibrium thermodynamics of a bistable chemical system: the schl ¨ogl model revisited

    Melissa Vellela and Hong Qian. Stochastic dynamics and non-equilibrium thermodynamics of a bistable chemical system: the schl ¨ogl model revisited. Journal of The Royal Society Interface, 6(39):925–940, 2009

    2009

  49. [57]

    Spectral representation and reduced order modeling of the dynamics of stochastic reaction networks via adaptive data partitioning

    Khachik Sargsyan, Bert Debusschere, Habib Najm, and Olivier Le Ma ˆıtre. Spectral representation and reduced order modeling of the dynamics of stochastic reaction networks via adaptive data partitioning. SIAM Journal on Scientific Computing, 31(6):4395–4421, 2010

    2010

  50. [58]

    Exact stochastic simulation of coupled chemical reactions

    Daniel T Gillespie. Exact stochastic simulation of coupled chemical reactions. The journal of physical chemistry, 81(25):2340–2361, 1977

    1977

  51. [59]

    Stochastic simulation of chemical kinetics

    Daniel T Gillespie. Stochastic simulation of chemical kinetics. Annu. Rev. Phys. Chem., 58(1):35–55, 2007

    2007

  52. [60]

    A neural ordinary differential equation framework for modeling inelastic stress response via internal state variables

    Reese E Jones, Ari L Frankel, and KL Johnson. A neural ordinary differential equation framework for modeling inelastic stress response via internal state variables. Journal of Machine Learning for Modeling and Computing, 3(3), 2022

    2022

  53. [61]

    A linear viscoelastic model calibration of sylgard

    Kevin Nicholas Long and Judith Alice Brown. A linear viscoelastic model calibration of sylgard

  54. [62]

    Measuring and testing dependence by correlation of distances

    G ´abor J Sz ´ekely, Maria L Rizzo, and Nail K Bakirov. Measuring and testing dependence by correlation of distances. The Annals of Statistics, 35(6):2769, 2007

    2007

  55. [63]

    Brownian distance covariance

    G ´abor J Sz ´ekely and Maria L Rizzo. Brownian distance covariance. The annals of applied statistics , pages 1236–1265, 2009

    2009

  56. [64]

    Optimal experimental design

    Valerii Fedorov. Optimal experimental design. Wiley Interdisciplinary Reviews: Computational Statistics , 2(5):581–589, 2010

    2010

  57. [65]

    Improv- ing the performance of stein variational inference through extreme sparsification of physically-constrained neural network models

    Govinda Anantha Padmanabha, Jan Niklas Fuhg, Cosmin Safta, Reese E Jones, and Nikolaos Bouklas. Improv- ing the performance of stein variational inference through extreme sparsification of physically-constrained neural network models. Computer Methods in Applied Mechanics and ...

    2024

  58. [66]

    Condensed stein variational gradient descent for uncertainty quantification of neural networks

    Govinda Anantha Padmanabha, Cosmin Safta, Nikolaos Bouklas, and Reese E Jones. Condensed stein variational gradient descent for uncertainty quantification of neural networks. arXiv preprint arXiv:2412.16462, 2024

    2024 arXiv

  59. [67]

    U-net: Convolutional networks for biomedical image seg- mentation

    Olaf Ronneberger, Philipp Fischer, and Thomas Brox. U-net: Convolutional networks for biomedical image seg- mentation. In Medical image computing and computer-assisted intervention–MICCAI 2015: 18th international conference, Munich, Germany, October 5-9, 2015, proceedings, par...

    2015

  60. [68]

    Freeu: Free lunch in diffusion u-net

    Chenyang Si, Ziqi Huang, Yuming Jiang, and Ziwei Liu. Freeu: Free lunch in diffusion u-net. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 4733–4743, 2024

    2024

  61. [69]

    Input convex neural networks

    Brandon Amos, Lei Xu, and J Zico Kolter. Input convex neural networks. InInternational conference on machine learning, pages 146–155. PMLR, 2017

    2017

  62. [70]

    JAX: composable transfor- mations of Python+NumPy programs, 2018

    James Bradbury, Roy Frostig, Peter Hawkins, Matthew James Johnson, Chris Leary, Dougal Maclaurin, George Necula, Adam Paszke, Jake VanderPlas, Skye Wanderman-Milne, and Qiao Zhang. JAX: composable transfor- mations of Python+NumPy programs, 2018

    2018

  63. [71]

    Equinox: neural networks in jax via callable pytrees and filtered transforma- tions

    Patrick Kidger and Cristian Garcia. Equinox: neural networks in jax via callable pytrees and filtered transforma- tions. arXiv preprint arXiv:2111.00254, 2021. 19 UQ of NN models via Langevin sampling A PREPRINT

    2021 arXiv

  64. [72]

    Gaussian process approximations of stochastic differential equations

    Cedric Archambeau, Dan Cornford, Manfred Opper, and John Shawe-Taylor. Gaussian process approximations of stochastic differential equations. In Gaussian Processes in Practice, pages 1–16. PMLR, 2007

    2007

  65. [73]

    Variational inference for diffusion processes

    C ´edric Archambeau, Manfred Opper, Yuan Shen, Dan Cornford, and John Shawe-Taylor. Variational inference for diffusion processes. Advances in neural information processing systems, 20, 2007

    2007

  66. [74]

    A Foundation in Digital Communication

    Amos Lapidoth. A Foundation in Digital Communication. Cambridge University Press, 2 edition, 2017. ‘ A The Kullback-Liebler divergence between two random processes Following Refs. [18, 72, 73] we provide a more step-by-step derivation of the joint Kullback-Liebler (KL) diverge...

    2017

  67. [184]

    https://www.osti.gov/biblio/1365535

    Technical report, Sandia National Lab.(SNL-NM), Albuquerque, NM (United States), 2017. https://www.osti.gov/biblio/1365535

    2017

Pith tools

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