Pith. sign in

REVIEW 3 major objections 7 minor 53 references

A Convex Approximation Framework for Neural Likelihood-Based Bayesian Inverse Problems

T0 review · 3 major / 7 minor · reviewed 2026-07-08 · glm-5.2

Pith's one-line read Folding normalization into the loss makes likelihood learning convex

desk verdict Convexity of KL objective for un-normalized neural likelihood approximation is clean and new; consistency proof is correct but limited to finite dimensions and unverified in experiments. read the letter →

arxiv 2607.06252 v1 pith:VGWXOAOR submitted 2026-07-07 stat.ML cs.LGmath.PRstat.ME

classification stat.MLcs.LGmath.PRstat.ME MSC 62G0565C0565N2165N7560B12
keywords neurallikelihoodapproximationBayesianinverseproblemsconvexoptimizationKLdivergenceun-normalizedmodelsconsistencyMonteCarloestimationposteriorsampling
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 addresses a core bottleneck in neural likelihood approximation for Bayesian inverse problems: standard approaches require the learned likelihood to be a properly normalized probability distribution, which forces the use of restrictive parametric families (like normalizing flows) that may not contain the true likelihood. The authors propose a different formulation: optimize directly over un-normalized negative log-likelihood functions, and absorb the normalization constant into the training objective itself. They show that the resulting objective—expected negative log-likelihood plus a log-normalizer term—is strictly convex in the un-normalized potential function. This means that, at the level of the function being learned (rather than the neural network parameters representing it), the optimization landscape has a unique global minimum. They further prove that when the objective is estimated from finite data using a nested Monte Carlo scheme, the minimizer converges in probability to the true likelihood as the sample size grows. The key identity enabling all of this is that the expected KL-divergence between the true posterior and the approximate posterior decomposes into a constant plus two terms—E[f(x,y)] and E[log Z_f(y)]—both computable from samples of the joint distribution without knowing the true likelihood.

What carries the argument

Strict convexity via Hölder inequality applied to log-partition function; nested Monte Carlo estimation of normalizer; bracketing entropy bounds for Hölder-class functions

What would settle it

If gradient-based training of neural network surrogates in this framework systematically converges to different local optima in function space (not just parameter space) for different random initializations, the practical value of the convexity guarantee would be undermined.

Watch

Extended reading notes

Core claim

The central object is the objective function Phi(f) = E_lambda[f(x;y)] + E_pi[log Z_f(y)], where f is an un-normalized negative log-likelihood and Z_f(y) is its normalizing integral. The paper proves this objective is strictly convex in f (Theorem 2.8), with the proof turning on a single application of Hölder's inequality to the log-normalizer term. This convexity holds in the function-space formulation: the set of normalized densities (exp(-f) integrating to one) is not convex, but the set of un-normalized potentials is, and the normalization is handled implicitly by the objective. The consistency result (Theorem 3.7) then follows by verifying two classical conditions—well-separated minimum

Load-bearing premise

The convexity guarantee holds in function space, but when the un-normalized potential is represented by a neural network, the optimization problem in network parameters remains non-convex. The practical benefit of the convexity result depends on whether gradient-based training can actually find the global optimum in parameter space, which the convexity result alone does not ensure.

Editorial extensions

If this is right

  • Convexity in function space provides an identifiability guarantee: even if neural network training finds different parameter settings, they must represent the same optimal likelihood function, which constrains the solution space.
  • The free-form approximation (learning f directly without assuming a forward map or noise model) can recover posteriors accurately even when the observational noise level is unknown, as demonstrated in the deblurring experiment.
  • The semiconductor doping profile experiment shows that neural likelihood surrogates can be approximately 200x faster than PDE-based forward model evaluations on CPU, enabling MCMC-based posterior sampling in minutes rather than days.
  • The framework extends naturally to residual and calibrated residual approximations where partial knowledge of the forward model or noise structure is available, with the free-form variant performing best when noise is unknown.

Reading between the lines

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

  • The gap between convexity in function space and non-convexity in parameter space suggests that architectural choices ensuring the parameter-to-function map is well-behaved could make the practical optimization problem more tractable, though this is not addressed in the paper.
  • The consistency proof is restricted to finite-dimensional H = R^n due to the reliance on classical bracketing entropy results for Hölder spaces on Euclidean domains; extending to infinite-dimensional Hilbert spaces would require new empirical process theory for function classes on abstract spaces.
  • The nested Monte Carlo structure of the estimator means the choice of inner sample size M(N) as a function of outer sample size N affects convergence; the paper proves convergence but does not optimize this trade-off, which could matter for practical efficiency.
  • The convexity result depends on the function class F being convex and closed; whether commonly used neural network architectures induce function classes that are approximately convex in the relevant sense is an empirical question left open.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 7 minor

Summary. This paper proposes a convex approximation framework for neural likelihood estimation in Bayesian inverse problems. The key idea is to work with un-normalized negative log-likelihood (NLL) potentials and fold normalization into the training objective, yielding a functional Φ(f) = E_λ[f(x;y)] + E_π[log Z_f(y)] that is strictly convex in f (Theorem 2.8). The authors establish consistency of empirical minimizers of a data-driven approximation Φ_N to the true potential L as the sample size grows (Theorem 3.7), under assumptions that include a sub-Gaussian prior, a convex closed function class F containing L with quadratic growth bounds, and finite-dimensional H = R^n with C^1 regularity for bracketing entropy arguments. Numerical experiments on a deblurring problem and a nonlinear PDE-based semiconductor doping inverse problem demonstrate the practical viability of the approach.

Significance. The paper makes a genuine theoretical contribution by identifying that the KL-based objective for likelihood approximation is convex in the un-normalized potential, a non-trivial observation that distinguishes the framework from normalizing-flow-based approaches where the normalized function class is not convex. The consistency result (Theorem 3.7) is a clean application of classical M-estimation theory (van der Vaart, Theorem 3.1) combined with bracketing entropy results from Nickl-Pötscher. The decomposition in Lemma 2.7 is verified and provides the foundation for the data-driven estimator. The numerical experiments are illustrative and include a practically relevant PDE-based inverse problem with significant computational speedups. The framework is transparent about its limitations, particularly the finite-dimensional restriction of the consistency proof.

major comments (3)
  1. §3.1, Assumption 3 and Theorem 3.7: The consistency guarantee requires H = R^n and F to be a bounded subset of C^1(R^{n+m}, -2), while the paper's motivation and formulation (§1.2, §2) emphasize infinite-dimensional Hilbert space H. The convexity results (Lemma 2.7, Theorem 2.8) are indeed stated in the infinite-dimensional setting, but the consistency result — which is the main theoretical contribution beyond convexity — only holds after discretization. The authors acknowledge this in §5, but the gap between the infinite-dimensional framing and the finite-dimensional theorem is load-bearing for the paper's claimed contribution to 'Bayesian inverse problems on function spaces.' The authors should either (a) clarify in the introduction and theorem statements that the consistency result is finite-dimensional and discuss what is needed for the infinite-dimensional extension, or (b) provide,
  2. §4, Remark 2.9: The convexity of Φ is in function space, but the numerical experiments use neural network parametrizations where optimization in parameter space remains non-convex. The paper does not assess whether gradient-based training actually finds the global optimum (or a near-optimal point) in the parameter space for the experiments in §4. Since the practical utility of the convexity result depends on this, some discussion of optimization quality — e.g., comparing solutions from multiple random initializations, or reporting training loss variability — would strengthen the connection between theory and practice. This is acknowledged in Remark 2.9 but not addressed experimentally.
  3. §4.2: The calibrated residual approximation fails to generate accurate posterior variance and means (L^1 bias of 1.072 vs. 0.1839 for free-form), but no diagnosis is offered. Since the calibrated residual is a natural variant of the proposed framework, understanding why it fails — whether due to optimization difficulty, insufficient expressivity, or the joint learning of forward map and noise covariance — would be valuable for practitioners. At minimum, the authors should state whether this failure is reproducible across initialization seeds.
minor comments (7)
  1. §2.1, Eq. (7): The lower bound involves 'for every ε > 0 there exists δ > 0' — the dependence of δ on ε is somewhat unusual and could be clarified with a brief remark on why this formulation is needed rather than a fixed pair (C_2^-, C_2^+).
  2. §3.2: The M(N) schedule is left unspecified ('we leave the choice of an optimal function M open'). While acceptable for a consistency result, a brief discussion of practical choices used in the experiments (what M(N) was used?) would help reproducibility.
  3. §4.1: The pCN step size β = 0.2 is stated without justification. A brief remark on acceptance rates or sensitivity to β would be helpful.
  4. Table 1: The GP regression column lists 'GP(x,y)' as output but the text describes a squared exponential covariance. Clarify whether the GP regresses the NLL directly or the log-likelihood.
  5. §4.2: Training details mention '30 epochs' but no learning rate, optimizer, or batch size is given. These should be specified for reproducibility.
  6. Reference [11] (Helin & Schneider) is listed as 'in preparation, 2026' — this should be updated to a citable form if available by the time of revision.
  7. §2.2, Definition 2.3: The notation F_Φ for the equivalence class is introduced but occasionally the paper writes F when F_Φ is meant (e.g., in Theorem 3.5, 'for any f ∈ F_Φ' vs. Theorem 3.6, 'sup_{f ∈ F}'). Consistency in notation would improve readability.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for a careful and constructive report. The referee correctly identifies the core contributions (convexity of the KL-based objective in the un-normalized potential, and the consistency result) and raises three substantive points: (1) the gap between infinite-dimensional framing and the finite-dimensional consistency theorem, (2) the lack of experimental evidence on optimization quality in neural network parameter space, and (3) the undiagnosed failure of the calibrated residual approximation in Section 4.2. We address each below and describe revisions we will make.

read point-by-point responses
  1. Referee: §3.1, Assumption 3 and Theorem 3.7: The consistency guarantee requires H = R^n and F to be a bounded subset of C^1(R^{n+m}, -2), while the paper's motivation and formulation (§1.2, §2) emphasize infinite-dimensional Hilbert space H. The gap between the infinite-dimensional framing and the finite-dimensional theorem is load-bearing for the paper's claimed contribution to 'Bayesian inverse problems on function spaces.' The authors should either (a) clarify in the introduction and theorem statements that the consistency result is finite-dimensional and discuss what is needed for the infinite-dimensional extension, or (b) provide the infinite-dimensional extension.

    Authors: The referee is correct that there is a genuine gap between the infinite-dimensional setting in which the convexity results (Lemma 2.7, Theorem 2.8) are formulated and the finite-dimensional setting required for the consistency theorem (Theorem 3.7). We do not dispute this. The root cause is that the bracketing entropy result we rely on (Theorem 3.9, from Nickl–Pötscher) is stated for bounded subsets of Hölder spaces on R^{n+m}, and extending this to function classes on infinite-dimensional domains is non-trivial and beyond what the current manuscript establishes. We will adopt option (a): we will revise the introduction (§1.2 and §1.3) and the theorem statements in §3 to state explicitly and prominently that the consistency result (Theorem 3.7) is finite-dimensional, while the convexity results hold in the infinite-dimensional setting. We will also add a dedicated discussion of what would be needed for the infinite-dimensional extension — specifically, an appropriate bracketing entropy theory for function classes on Hilbert-space-valued domains, which is an active area of research (e.g., Nickl's recent work on Bayesian non-linear statistical inverse problems). We agree with the referee that the current framing could mislead readers into thinking the consistency result applies directly to function spaces, and this will be corrected. revision: yes

  2. Referee: §4, Remark 2.9: The convexity of Φ is in function space, but the numerical experiments use neural network parametrizations where optimization in parameter space remains non-convex. The paper does not assess whether gradient-based training actually finds the global optimum (or a near-optimal point) in the parameter space for the experiments in §4. Some discussion of optimization quality — e.g., comparing solutions from multiple random initializations, or reporting training loss variability — would strengthen the connection between theory and practice.

    Authors: The referee raises a fair point. Remark 2.9 acknowledges that convexity in function space does not imply convexity in parameter space, but the experiments do not provide evidence about whether gradient-based optimization reliably finds near-optimal solutions. We will address this by adding experiments comparing training loss and posterior quality across multiple random initializations for the deblurring problem (§4.1). Specifically, we will report the variability in the final training objective and in the resulting posterior L^1 bias across at least 5 independent random seeds. This will give practitioners a concrete sense of optimization stability. We note that the convexity result still provides value through the identifiability guarantee (uniqueness of the population minimizer in F_Φ), which underlies the consistency analysis, even if the parameter-space optimization landscape is non-convex. We will make this point more explicit in the revised Remark 2.9. revision: yes

  3. Referee: §4.2: The calibrated residual approximation fails to generate accurate posterior variance and means (L^1 bias of 1.072 vs. 0.1839 for free-form), but no diagnosis is offered. Understanding why it fails — whether due to optimization difficulty, insufficient expressivity, or the joint learning of forward map and noise covariance — would be valuable for practitioners. At minimum, the authors should state whether this failure is reproducible across initialization seeds.

    Authors: We agree that the failure of the calibrated residual approximation in §4.2 is insufficiently diagnosed. Based on our experience with the experiments, we believe the most likely cause is the joint learning of the forward map and the noise covariance: the additional diagonal layer estimating Γ introduces a non-identifiability between the forward map error and the noise level, which can destabilize training. However, we have not yet systematically verified this hypothesis or confirmed reproducibility across seeds. We will add experiments reporting the calibrated residual results across multiple initialization seeds to determine whether the failure is consistent. We will also add a diagnostic comparison: training the calibrated residual with the noise covariance fixed to the true value (which reduces it to the residual approximation) versus learned, to isolate whether the joint estimation of Γ is the source of the problem. If time constraints prevent completing the full diagnostic, we will at minimum state clearly that the failure is reproducible across seeds and offer our best assessment of the cause, while being transparent that a complete diagnosis is left for future work. revision: partial

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity; derivation chain is self-contained with standard tools.

full rationale

The paper's two main results are derived from first principles using standard mathematical tools, not from self-citation chains. (1) The key identity (Lemma 2.7) follows from direct algebraic manipulation of the KL divergence definition — no circularity. (2) Strict convexity (Theorem 2.8) is proved via Hölder's inequality applied to the log-partition function, with the equality condition giving identifiability — a self-contained argument. (3) Consistency (Theorem 3.7) combines two sub-results: Theorem 3.5 (well-separated minimum) follows from the non-negativity of KL divergence via Lemma 2.7, and Theorem 3.6 (uniform LLN) uses bracketing entropy from Nickl-Pötscher [29], an external result. The overarching consistency framework (Theorem 3.1) is from van der Vaart [47], also external. The self-citation to [11] (Helin & Schneider, 'in preparation') is explicitly described as motivation ('Our work is motivated by [11]') and does not serve as a load-bearing premise for any proof — Theorem 2.8's proof stands independently. Reference [10] (co-authored by Helin) is cited for the assumption framework and a sufficient-conditions lemma (Example 2.1), but the main theorems do not depend on it. No 'prediction' reduces to a fitted input or a self-cited result by construction. The score of 1 reflects the minor motivational self-citation to [11], which is not load-bearing.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

No new physical entities, particles, or forces are introduced. The framework operates within standard Bayesian inverse problem machinery. The free parameters are experimental hyperparameters, not theoretical constants fitted to data and presented as predictions.

free parameters (3)
  • pCN step size β = 0.2
    Chosen for the MCMC sampler in the deblurring experiment, Section 4.1.
  • Neural network parameters θ = unspecified count
    Learnable parameters for the NLL approximations; architecture described but exact parameter counts and training hyperparameters not fully specified.
  • M(N) schedule = unspecified
    The number of inner-loop prior samples for the nested Monte Carlo estimator; stated as coercive function M: N→N but the specific choice used in experiments is not given.
assumptions (4)
  • domain assumption Prior satisfies E_μ exp(C₁||x||²) < ∞ (Assumption 1)
    Sub-Gaussian prior requirement; standard for Bayesian inverse problems on Hilbert spaces, invoked throughout Section 2.
  • domain assumption True potential L ∈ F where F is convex, closed, non-negative (Assumption 2)
    Load-bearing: convexity of Φ is useless without convexity of F. Invoked in Theorems 2.8, 3.5, 3.6, 3.7.
  • domain assumption F is bounded subset of C¹(R^{n+m}, -2) (Assumption 3)
    Required for finite bracketing entropy via Theorem 3.9 (Nickl-Pötscher). Restricts consistency to finite-dimensional H=R^n.
  • domain assumption Quadratic growth bounds on f (Eq. 7)
    Ensures Z_f(y) is well-defined and finite; invoked in Lemma A.1 and throughout.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Convex Approximation Framework for Neural Likelihood-Based Bayesian Inverse Problems." pith.science (2026). https://pith.science/paper/VGWXOAOR

@misc{pith2026260706252,
  author       = {Pith},
  title        = {Pith review of: A Convex Approximation Framework for Neural Likelihood-Based Bayesian Inverse Problems},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VGWXOAOR}},
  note         = {Machine review of arXiv:2607.06252}
}
read the original abstract

Many problems in science and engineering are difficult to model accurately, either due to unknown physical mechanisms, poorly quantified measurement uncertainty, or prohibitive computational costs of high-fidelity simulations. These challenges limit the applicability of classical probabilistic inference methods such as Markov chain Monte Carlo, especially in high-dimensional Bayesian inverse problems. As data from scientific experiments become increasingly available, machine learning methods offer a flexible alternative to explicit parametric modelling. We study neural likelihood approximation, where the goal is to learn the likelihood function directly from data without explicit knowledge of the underlying data-generating process. A common approach trains likelihood surrogates by minimizing the Kullback-Leibler divergence between the true posterior and an approximate posterior, which is equivalent to minimizing the expected negative log-likelihood. This work improves the theoretical foundations of neural likelihood approximation by alleviating limitations of restrictive model classes: we show that, by working with un-normalized potentials and folding normalization into the training objective, the resulting learning problem is strictly convex. We show that empirical minimizers of the resulting data-driven objective converge to the true likelihood as the sample size grows. Numerical experiments for the neural likelihood approximation are conducted for a deblurring and a non-linear PDE based imaging problem.

Figures

Figures reproduced from arXiv: 2607.06252 by the authors.

Figure 1
Figure 1. Posterior MCMC samples for the deblurring problem. From left to right: true posterior, free-form approximation, residual approximation and cali￾brated residual approximation. All methods use only 1, 000 training points [PITH_FULL_IMAGE:figures/full_fig_p011_1.png] view at source ↗
Figure 2
Figure 2. Posterior MCMC samples for the deblurring problem. From left to right: true posterior, GP regression with 1,000 training points and GP regression for the likelihood with 10,000 training points. data-efficient way to accurately approximate the likelihood in inverse problems. In settings where the observational noise is unknown, the free-form approximation seems preferable over a calibrated residual approximation base… view at source ↗
Figure 3
Figure 3. Summary of MCMC samples. Top row: True quantity of interest followed by posterior mean (left), bias (middle) and posterior variance (right) of the posterior samples with the true PDE-based NLL and its approximations. The L 1 norm of the biases are given by 0.1807 (exact), 0.1839 (free-form), 0.1954 (residual), 1.072 (calibrated residual). 5. Conclusions This work developed a theoretical foundation for neural likelih… view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

53 extracted references · 53 canonical work pages

  1. [1]

    Approximation errors and model reduction with an application in optical diffusion tomography

    S R Arridge et al. “Approximation errors and model reduction with an application in optical diffusion tomography”. In:Inverse Problems22.1 (Jan. 2006), p. 175.doi:10.1088/0266- 5611/22/1/010

  2. [2]

    Optical tomography in medical imaging

    Simon R Arridge. “Optical tomography in medical imaging”. In:Inverse Problems15.2 (1999), R41

  3. [3]

    Risks for the Long Run: A Potential Resolution of Asset Pricing Puzzles

    Ravi Bansal and Amir Yaron. “Risks for the Long Run: A Potential Resolution of Asset Pricing Puzzles”. In:The Journal of Finance59.4 (2004), pp. 1481–1509.doi:https : //doi.org/10.1111/j.1540-6261.2004.00670.x

  4. [4]

    Arved Bartuska et al.Double-loop randomized quasi-Monte Carlo estimator for nested inte- gration. 2025. arXiv:2302.14119 [math.NA].url:https://arxiv.org/abs/2302.14119

  5. [5]

    Nada Cvetkovi´ c and Han Cheng Lie.Upper and lower bounds for local Lipschitz stability of Bayesian posteriors. 2025. arXiv:2505.23541 [math.ST].url:https://arxiv.org/abs/ 2505.23541

  6. [6]

    Truncated proposals for scalable and hassle-free simulation-based inference

    Michael Deistler, Pedro J Goncalves, and Jakob H Macke. “Truncated proposals for scalable and hassle-free simulation-based inference”. In:Advances in Neural Information Processing Systems. Ed. by S. Koyejo et al. Vol. 35. Curran Associates, Inc., 2022, pp. 23135–23149

  7. [7]

    Deep Surrogate Accelerated Delayed- Acceptance Hamiltonian Monte Carlo for Bayesian Inference of Spatio-Temporal Heat Fluxes in Rotating Disc Systems

    Teo Deveney, Eike H. Mueller, and Tony Shardlow. “Deep Surrogate Accelerated Delayed- Acceptance Hamiltonian Monte Carlo for Bayesian Inference of Spatio-Temporal Heat Fluxes in Rotating Disc Systems”. In:SIAM/ASA Journal on Uncertainty Quantification11.3 (2023), pp. 970–995.doi:10.1137/22M1513113

  8. [8]

    The Statistical Accuracy of Neural Posterior and Likelihood Estimation

    David T. Frazier et al.The Statistical Accuracy of Neural Posterior and Likelihood Estima- tion. 2024. arXiv:2411.12068 [stat.ML].url:https://arxiv.org/abs/2411.12068

Show all 53 references
  1. [9]

    Bayesian Optimization for Likelihood-Free In- ference of Simulator-Based Statistical Models

    Michael U. Gutmann and Jukka Corander. “Bayesian Optimization for Likelihood-Free In- ference of Simulator-Based Statistical Models”. In:Journal of Machine Learning Research 17.125 (2016), pp. 1–47.url:http://jmlr.org/papers/v17/15-017.html

  2. [10]

    Tapio Helin, Youssef Marzouk, and Jose Rodrigo Rojo-Garcia.Bayesian optimal experimen- tal design with Wasserstein information criteria. 2025. arXiv:2504.10092 [stat.ME].url: https://arxiv.org/abs/2504.10092

  3. [11]

    In preparation

    Tapio Helin and Fabian Schneider.Optimal approximation error approach in Bayesian in- verse problems. In preparation. 2026

  4. [12]

    Introduction to Gaussian Process Regression in Bayesian Inverse Prob- lems, with New Results on Experimental Design for Weighted Error Measures

    Tapio Helin et al. “Introduction to Gaussian Process Regression in Bayesian Inverse Prob- lems, with New Results on Experimental Design for Weighted Error Measures”. In: July 2024, pp. 49–79.isbn: 978-3-031-59761-9.doi:10.1007/978-3-031-59762-6_3

  5. [13]

    Deep neural network expression of posterior expectations in Bayesian PDE inversion

    Lukas Herrmann, Christoph Schwab, and Jakob Zech. “Deep neural network expression of posterior expectations in Bayesian PDE inversion”. In:Inverse Problems36.12 (Dec. 2020), p. 125011.doi:10.1088/1361-6420/abaf64

  6. [14]

    Yuga Hikida et al.Multilevel neural simulation-based inference. 2025. arXiv:2506.06087 [stat.ML].url:https://arxiv.org/abs/2506.06087

  7. [15]

    Electrical Impedance Tomography: Methods, History and Applications

    D.S. Holder. “Electrical Impedance Tomography: Methods, History and Applications”. In: vol. 32. Jan. 2005.isbn: 0750309520

  8. [16]

    Kaipio and E

    J. Kaipio and E. Somersalo.Statistical and Computational Inverse Problems. Applied Math- ematical Sciences. Springer New York, 2006.isbn: 9780387271323

  9. [17]

    Simple and Scalable Predictive Uncertainty Estimation using Deep Ensembles

    Balaji Lakshminarayanan, Alexander Pritzel, and Charles Blundell. “Simple and Scalable Predictive Uncertainty Estimation using Deep Ensembles”. In:Advances in Neural Informa- tion Processing Systems. Ed. by I. Guyon et al. Vol. 30. Curran Associates, Inc., 2017

  10. [18]

    Discretization-invariant Bayesian inver- sion and Besov space priors

    Matti Lassas, Eero Saksman, and Samuli Siltanen. “Discretization-invariant Bayesian inver- sion and Besov space priors”. In:Inverse Problems and Imaging3.1 (2009), pp. 87–122.issn: 1930-8337.doi:10.3934/ipi.2009.3.87

  11. [19]

    Linear inverse problems for gen- eralised random variables

    Markku Lehtinen, Lassi P¨ aiv¨ arinta, and Erkki Somersalo. “Linear inverse problems for gen- eralised random variables”. English. In:Inverse Problems5.4 (1989), pp. 599–612.issn: 0266-5611.doi:10.1088/0266-5611/5/4/011. REFERENCES 21

  12. [20]

    Surrogate modeling for Bayesian inverse prob- lems based on physics-informed neural networks

    Yongchao Li, Yanyan Wang, and Liang Yan. “Surrogate modeling for Bayesian inverse prob- lems based on physics-informed neural networks”. In:Journal of Computational Physics475 (2023), p. 111841.issn: 0021-9991.doi:https://doi.org/10.1016/j.jcp.2022.111841

  13. [21]

    Random Forward Models and Log- Likelihoods in Bayesian Inverse Problems

    H. C. Lie, T. J. Sullivan, and A. L. Teckentrup. “Random Forward Models and Log- Likelihoods in Bayesian Inverse Problems”. In:SIAM/ASA Journal on Uncertainty Quan- tification6.4 (2018), pp. 1600–1629.doi:10.1137/18M1166523

  14. [22]

    Likelihood-free inference with emulator networks

    Jan-Matthis Lueckmann et al. “Likelihood-free inference with emulator networks”. In:Pro- ceedings of The 1st Symposium on Advances in Approximate Bayesian Inference. Ed. by Francisco Ruiz et al. Vol. 96. Proceedings of Machine Learning Research. PMLR, Dec. 2019, pp. 32–53

  15. [23]

    Bayesian synthetic likelihood for stochastic models with applications in mathematical finance

    Ramona Maraia et al. “Bayesian synthetic likelihood for stochastic models with applications in mathematical finance”. In:Frontiers in Applied Mathematics and StatisticsVolume 9 - 2023 (2023).issn: 2297-4687.doi:10.3389/fams.2023.1187878

  16. [24]

    Approximate Bayesian computational methods

    Jean-Michel Marin et al. “Approximate Bayesian computational methods”. In:Statistics and Computing22.6 (Nov. 1, 2012), pp. 1167–1180.issn: 1573-1375.doi:10.1007/s11222-011- 9288-2

  17. [25]

    A Stochastic Collocation Approach to Bayesian Infer- ence in Inverse Problems

    Youssef Marzouk and Dongbin Xiu. “A Stochastic Collocation Approach to Bayesian Infer- ence in Inverse Problems”. In:PRISM: NNSA Center for Prediction of Reliability, Integrity and Survivability of Microsystems6 (Oct. 2009).doi:10.4208/cicp.2009.v6.p826

  18. [26]

    Dimensionality reduction and polynomial chaos acceleration of Bayesian inference in inverse problems

    Youssef M. Marzouk and Habib N. Najm. “Dimensionality reduction and polynomial chaos acceleration of Bayesian inference in inverse problems”. In:Journal of Computational Physics 228.6 (2009), pp. 1862–1902.issn: 0021-9991

  19. [27]

    GPS-ABC: Gaussian process surrogate approximate Bayesian computation

    Edward Meeds and Max Welling. “GPS-ABC: Gaussian process surrogate approximate Bayesian computation”. In:Proceedings of the Thirtieth Conference on Uncertainty in Ar- tificial Intelligence. UAI’14. Quebec City, Quebec, Canada: AUAI Press, 2014, pp. 593–602

  20. [28]

    Richard Nickl.Bayesian non-linear statistical inverse problems. eng. Zurich lectures in ad- vanced mathematics. Berlin: EMS Press, 2023.isbn: 3985470537

  21. [29]

    Bracketing Metric Entropy Rates and Empirical Central Limit Theorems for Function Classes of Besov- and Sobolev-Type

    Richard Nickl and Benedikt P¨ otscher. “Bracketing Metric Entropy Rates and Empirical Central Limit Theorems for Function Classes of Besov- and Sobolev-Type”. In:Journal of Theoretical Probability20 (June 2007), pp. 177–199.doi:10.1007/s10959-007-0058-1

  22. [30]

    Estimating the Transmission Dynamics of Streptococcus pneumo- niae from Strain Prevalence Data

    Elina Numminen et al. “Estimating the Transmission Dynamics of Streptococcus pneumo- niae from Strain Prevalence Data”. In:Biometrics69.3 (2013), pp. 748–757. (Visited on 11/19/2025)

  23. [31]

    George Papamakarios.Neural Density Estimation and Likelihood-free Inference. 2019. arXiv: 1910.13233 [stat.ML].url:https://arxiv.org/abs/1910.13233

  24. [32]

    Fast epsilon-free Inference of Simulation Models with Bayesian Conditional Density Estimation

    George Papamakarios and Iain Murray. “Fast epsilon-free Inference of Simulation Models with Bayesian Conditional Density Estimation”. In:Advances in Neural Information Pro- cessing Systems. Ed. by D. Lee et al. Vol. 29. Curran Associates, Inc., 2016

  25. [33]

    Masked Autoregressive Flow for Density Estimation

    George Papamakarios, Theo Pavlakou, and Iain Murray. “Masked Autoregressive Flow for Density Estimation”. In:Advances in Neural Information Processing Systems. Ed. by I. Guyon et al. Vol. 30. Curran Associates, Inc., 2017

  26. [34]

    Sequential Neural Likelihood: Fast Likelihood-free Inference with Autoregressive Flows

    George Papamakarios, David Sterratt, and Iain Murray. “Sequential Neural Likelihood: Fast Likelihood-free Inference with Autoregressive Flows”. In:Proceedings of the Twenty-Second International Conference on Artificial Intelligence and Statistics. Ed. by Kamalika Chaud- huri a...

  27. [35]

    Radev et al.BayesFlow: Learning complex stochastic models with invertible neural networks

    Stefan T. Radev et al.BayesFlow: Learning complex stochastic models with invertible neural networks. 2020. arXiv:2003.06281 [stat.ML].url:https://arxiv.org/abs/2003.06281

  28. [36]

    On Nesting Monte Carlo Estimators

    Tom Rainforth et al. “On Nesting Monte Carlo Estimators”. In:Proceedings of the 35th International Conference on Machine Learning. Ed. by Jennifer Dy and Andreas Krause. Vol. 80. Proceedings of Machine Learning Research. PMLR, June 2018, pp. 4267–4276

  29. [37]

    Using Likelihood-Free Inference to Compare Evolutionary Dynamics of the Protein Networks of H. pylori and P. falciparum

    Oliver Ratmann et al. “Using Likelihood-Free Inference to Compare Evolutionary Dynamics of the Protein Networks of H. pylori and P. falciparum”. In:PLOS Computational Biology 3.11 (Nov. 2007), pp. 1–13.doi:10.1371/journal.pcbi.0030230. 22 REFERENCES

  30. [38]

    Conditional Density Estimation with Neural Networks: Best Practices and Benchmarks

    Jonas Rothfuss et al. “Conditional Density Estimation with Neural Networks: Best Practices and Benchmarks”. In:arXiv:1903.00954(2019)

  31. [39]

    Bayesianly Justifiable and Relevant Frequency Calculations for the Ap- plied Statistician

    Donald B. Rubin. “Bayesianly Justifiable and Relevant Frequency Calculations for the Ap- plied Statistician”. In:The Annals of Statistics12.4 (1984), pp. 1151–1172.doi:10.1214/ aos/1176346785

  32. [40]

    Lec- ture notes

    Bodhisattva Sen.A Gentle Introduction to Empirical Process Theory and Applications. Lec- ture notes. July 2022

  33. [41]

    Sequential Monte Carlo without likelihoods

    S. A. Sisson, Y. Fan, and Mark M. Tanaka. “Sequential Monte Carlo without likelihoods”. In:Proceedings of the National Academy of Sciences104.6 (2007), pp. 1760–1765.doi:10. 1073/pnas.0607208104

  34. [42]

    On the local Lipschitz stability of Bayesian inverse problems

    Bj¨ orn Sprungk. “On the local Lipschitz stability of Bayesian inverse problems”. In:Inverse Problems36.5 (Apr. 2020), p. 055015.doi:10.1088/1361-6420/ab6f43

  35. [43]

    Inverse problems: A Bayesian perspective

    A. M. Stuart. “Inverse problems: A Bayesian perspective”. In:Acta Numerica19 (2010), pp. 451–559.doi:10.1017/S0962492910000061

  36. [44]

    Posterior consistency for Gaussian process approximations of Bayesian posterior distributions

    Andrew M. Stuart and Aretha L. Teckentrup. “Posterior consistency for Gaussian process approximations of Bayesian posterior distributions”. In:Mathematics of Computation87.310 (Mar. 2018).doi:10.1090/mcom/3244

  37. [45]

    Bayesian Inversion for the Identification of the Doping Profile in Unipolar Semiconductor Devices

    Leila Taghizadeh and Ansgar J¨ ungel. “Bayesian Inversion for the Identification of the Doping Profile in Unipolar Semiconductor Devices”. In:SIAM Journal on Scientific Computing47.3 (2025), B690–B709.doi:10.1137/24M1687042

  38. [46]

    van der Vaart and J.A

    A. van der Vaart and J.A. Wellner.Weak Convergence and Empirical Processes: With Ap- plications to Statistics. Springer Series in Statistics. Springer, 1996

  39. [47]

    A. W. van der Vaart.Asymptotic Statistics. Cambridge Series in Statistical and Probabilistic Mathematics. Cambridge University Press, 1998

  40. [48]

    Neural likelihood surfaces for spatial processes with computationally intensive or intractable likelihoods

    Julia Walchessen, Amanda Lenzi, and Mikael Kuusela. “Neural likelihood surfaces for spatial processes with computationally intensive or intractable likelihoods”. In:Spatial Statistics62 (2024), p. 100848.issn: 2211-6753.doi:https : / / doi . org / 10 . 1016 / j . spasta . 2024...

  41. [49]

    Hongji Wang et al.Sequential Bayesian Design for Efficient Surrogate Construction in the Inversion of Darcy Flows. 2025. arXiv:2507.17713 [stat.ML].url:https://arxiv.org/ abs/2507.17713

  42. [50]

    Statistical inference for noisy nonlinear ecological dynamic systems

    Simon N. Wood. “Statistical inference for noisy nonlinear ecological dynamic systems”. In: Nature466.7310 (Aug. 2010), pp. 1102–1104.issn: 1476-4687.doi:10.1038/nature09319

  43. [51]

    Princeton University Press, 2010

    Dongbin Xiu.Numerical Methods for Stochastic Computations: A Spectral Method Approach. Princeton University Press, 2010

  44. [52]

    An adaptive surrogate modeling based on deep neural networks for large-scale Bayesian inverse problems

    Liang Yan and Tao Zhou. “An adaptive surrogate modeling based on deep neural networks for large-scale Bayesian inverse problems”. In:Communications in Computational Physics 28.5 (Nov. 2020), pp. 2180–2205

  45. [53]

    Solving high-dimensional inverse problems using amortized likelihood-free inference with noisy and incomplete data

    Jice Zeng et al. “Solving high-dimensional inverse problems using amortized likelihood-free inference with noisy and incomplete data”. In:Computer Methods in Applied Mechanics and Engineering443 (2025), p. 118064.issn: 0045-7825.doi:https://doi.org/10.1016/j. cma.2025.118064

Pith tools

Reviewed July 8, 2026 · model on record in the stance chip above.