Pith. sign in

REVIEW 3 major objections 5 minor 56 references

Efficient Bayesian Optimal Experimental Design for Expensive Computational Models over Finite Design Sets

T0 review · 3 major / 5 minor · reviewed 2026-08-01 · deepseek-v4-flash

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

Pith's one-line read An adaptive elimination algorithm for Bayesian optimal experimental design identifies the best experiment in a finite set with orders of magnitude fewer expensive forward-model evaluations than standard nested Monte Carlo, without sacrifici

desk verdict Solid assembly of known tricks with two genuinely useful estimator variants (LRNMC, ANMC); the reliability claim leans on an unproven monotone-bias assumption, but the paper deserves serious refereeing. read the letter →

arxiv 2607.16933 v1 pith:HHX4PGYK submitted 2026-07-18 cs.CE

Maximilian Dinkel , Dragos C. Ana , Benedikt Goderbauer , Wolfgang A. Wall This is my paper
classification cs.CE MSC 62K0562F1565C05
keywords BayesianoptimalexperimentaldesignfinitesetsnestedMonteCarloadaptiveeliminationRao-Blackwellizationcommonrandomnumbersbootstrapconfidenceexpectedinformationgain
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

Bayesian optimal experimental design usually forces a choice between expensive nested Monte Carlo estimates and biased local approximations. This paper argues that when the candidate designs form a finite set, the goal should not be to estimate information gain accurately but to compare designs reliably. It introduces an adaptive elimination algorithm that reuses parameter samples across the inner and outer Monte Carlo loops, excludes the current sample to keep the bias benign, averages over noise via Rao-Blackwellization, and uses common random numbers plus bootstrap confidence to delete inferior designs early. The result is a reduction in required model evaluations of about two orders of magnitude relative to a fixed-budget reuse baseline, while the optimal design is identified in essentially all test runs. If this holds for a target application, BOED becomes practical for expensive finite-element and multiphysics simulators.

What carries the argument

The central object is the ANMC estimator: it uses a single prior sample set for both the outer expectation and the inner marginal-likelihood estimate, omits the current index in the inner sum, and averages pointwise mutual information over multiple noise samples per outer sample. This makes each design's EIG estimate cheap, low-variance, and free of the negative sample-reuse bias. On top of it sits the adaptive design elimination loop, which shares one sample set across all designs (common random numbers), computes bootstrap probabilities of pairwise superiority, and eliminates a design only after the probability that it is worse than the current best exceeds 0.99 for three consecutive itera

What would settle it

Repeat the bias experiment of Figure 1 with a problem where the likelihood terms are less variable for higher-EIG designs, so the Jensen gap does not grow with EIG; if the ANMC bias is then flat or decreasing in true EIG, run the ADE algorithm on a known finite design set and check whether the true best design is ever eliminated with bootstrap probability above 0.99 and never recovered by later iterations.

Watch

Extended reading notes

Core claim

The paper's central claim is that the best experiment from a finite, unordered set can be identified without accurately estimating every expected information gain. Instead, the algorithm pairs an Accelerated Nested Monte Carlo (ANMC) estimator—which reuses outer-loop parameter samples in the inner marginal-likelihood sum, leaves out the current sample to remove a negative reuse bias, and Rao-Blackwellizes over noise draws—with a bootstrap-based elimination loop that deletes designs only after three consecutive confident inferiority verdicts. Because common random numbers across designs make the pairwise information-gain differences low-variance, and because the ANMC bias tends to grow with t

Load-bearing premise

The algorithm's reliability hinges on the empirical premise that the ANMC estimator's bias increases monotonically with the true expected information gain; if this 'facilitating bias' fails, the bootstrap confidence can certify a biased comparison and eliminate the true best design before more samples correct the ranking.

Editorial extensions

If this is right

  • BOED over finite design sets becomes feasible for expensive simulators: the ADE+ANMC combination identifies the optimal design with on the order of 10^2 to 10^3 model evaluations in the tested examples, where fixed-budget nested Monte Carlo would need 10^5 to 10^6.
  • The leave-out correction is a necessary ingredient: reusing the current outer sample in the inner loop adds a negative bias that compresses EIG differences and can make a worse design look competitive.
  • The algorithm is estimator-agnostic and black-box friendly: it requires only likelihood evaluations, not gradients or Gaussian approximations, and it works with non-Gaussian observation models.
  • The default hyperparameters (8 new samples per iteration, 0.99 probability threshold, 3 consecutive eliminations, 10^5 bootstrap samples) transfer across the four examples, so a practitioner can run ADE without tuning.

Reading between the lines

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

  • The paper leaves implicit that the efficiency gains are empirical rather than proven; a natural extension is a sample-complexity bound for the elimination rule under the monotone-bias premise, which would tell practitioners when the two-order-of-magnitude saving is guaranteed.
  • The bootstrap elimination threshold and persist count behave like a fixed-confidence racing procedure; connecting them to an error budget could turn the method into a procedure with a formal reliability guarantee.
  • If a target application violates the monotone-bias premise, a safer variant could combine ADE with a debiased or multilevel estimator, accepting more evaluations in exchange for a guarantee that the true best design is never eliminated early.
  • The method currently assumes a deterministic forward model with stochastic noise; extending the reuse and leave-out logic to stochastic simulators would open the same efficiency gains to a wider class of applications.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper proposes an adaptive design elimination (ADE) algorithm for Bayesian optimal experimental design over finite design sets. The method combines nested Monte Carlo estimation with outer-sample reuse, a leave-one-out correction (LRNMC), Rao–Blackwellization (ANMC), common random numbers across designs, and bootstrap-based elimination of inferior designs. The central claim is that ADE with ANMC identifies the optimal design with high reliability while reducing the number of forward model evaluations by orders of magnitude relative to static design comparison (SDC). The method is demonstrated on four examples: an analytic lognormal test case, a pharmacokinetic model, a human lung model, and a viscoplastic finite-element model.

Significance. If the claims hold, the paper is a useful practical contribution. It packages several known variance-reduction ideas into an automatic elimination procedure with sensible defaults, and the analytic example provides a clean correctness check. The code is publicly available. However, the central reliability claim rests on an empirical monotone-bias assumption that is not theoretically justified, and the expensive-model validation uses references from the same biased estimator family. These issues must be addressed before the method can be recommended as 'high reliability' for general large-scale applications.

major comments (3)
  1. [§2.3, Eq. (16), Algorithm 1] The reliability of ADE depends on the assertion that the bias of NMC/LRNMC/ANMC increases monotonically with the true EIG, thereby amplifying true differences between designs. This is demonstrated for one lognormal problem (Figure 1) and justified heuristically via Jensen's inequality. No theoretical analysis or broader evidence is provided. If the bias is not monotone, Eq. (16) can report high confidence in a wrong ordering, and Algorithm 1 may permanently eliminate the optimal design before later iterations correct the ranking, since eliminated designs are never revisited. This is a load-bearing assumption for the central reliability claim. The authors should either prove monotonicity under stated conditions, provide a robust elimination criterion that does not rely on it, or analyze the consequences when it fails.
  2. [§3.4, Table 4 and Figure 9] The reference optimal design ξ9 is identified using RNMC and ANMC at N=10^3 samples. At this sample size the two estimators disagree dramatically (e.g., for ξ9, RNMC gives 6.662 and ANMC gives 26.185). This indicates that neither estimator has converged and that the 'true' ranking is uncertain. The reported 100% accuracy of ADE in this example is therefore only agreement with the same biased estimator family, not with an independent ground truth. Since this is the only example with an expensive forward model, the claim that the method is 'well-suited for large-scale engineering applications' is not fully supported. Please provide a more reliable reference (e.g., larger N for at least a subset of designs or an independent estimator) or qualify the conclusion.
  3. [§2.4, Algorithm 1] The elimination rule has no formal error control. The bootstrap probability (16) is not a calibrated p-value, and the thresholds q=0.99 and NC=3 are heuristic. At early iterations with N=8, the bootstrap is based on very few samples and can be highly discrete; the paper does not analyze the probability of incorrectly eliminating the optimal design. The hyperparameter study in Appendix C is useful but empirical. For a method whose central promise is reliability, the authors should provide a finite-sample or asymptotic analysis of the elimination decision, or at least a conservative correction to the bootstrap threshold.
minor comments (5)
  1. [Algorithm 1] The line 'Identify reference design: k* = argmax_k P^(k)' is ambiguous because P^(k) is defined as a vector of per-sample PMI estimates. It should read 'argmax_k (1/|Θ|) Σ_i P_i^(k)' or otherwise clarify that the mean PMI is used.
  2. [§3.1, Figure 2] The claim of 'two orders of magnitude' reduction in model evaluations would be easier to verify if the figure included explicit axis markers or tabulated cost values at the 100% accuracy points.
  3. [§3.4, Table 4] Given the large discrepancy between RNMC and ANMC reference values, reporting standard errors or confidence intervals for the N=10^3 estimates would help the reader assess convergence.
  4. [§3.3] The statement that 'runs are not fully independent' because samples are drawn from a precomputed set could affect the reported accuracy values. Please quantify the dependence or sample fresh inputs for at least a subset of runs.
  5. [§2.2] The term 'accelerated' in ANMC is not explicitly justified; consider briefly noting that the acceleration comes from avoiding repeated model evaluations via Rao–Blackwellization.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the estimator chain is self-contained, with only a same-estimator validation caveat in §3.2–3.4.

full rationale

The claimed derivation is self-contained: EIG (5) feeds NMC (7), RNMC (8), LRNMC (9), ANMC (12), the CRN difference estimator (14), bootstrap probability (16), and the elimination rule in Algorithm 1. No parameter is fitted to reproduce a known answer, and the hyperparameters (Nnew=8, q=0.99, NC=3, NB=10^5, Nϵ=100) are fixed defaults rather than values calibrated to force the reported accuracy. The load-bearing premise in §2.3, the monotone-bias property, is explicitly empirical ('Empirically, we observe that in most settings the bias ... increases with the true EIG') and is not derived from the target result; if it fails in an application, ADE could eliminate the true best design, but that is a correctness risk rather than circularity. The validation in §3.2–3.4 does define the 'true' optimal design using the same NMC-family estimators that drive ADE (e.g., 'We approximate the true EIG values of all designs using the RNMC and the LRNMC estimator with N = 10^6 samples'), so those benchmarks measure consistency between the adaptive and fixed-sample versions of the same estimator rather than agreement with independent ground truth. This weakens external validity, but it is not circular by construction: the ADE decision is a random-subsample argmax and never uses the reference values. Section 3.1 additionally validates against a closed-form analytic EIG (Eq. 19), providing an independent external benchmark. The self-citations ([34,35,47,50]) are methodological or software citations and are not load-bearing in the derivation. Overall, no step reduces to its own inputs; the paper's central claim has independent content, and any concern about the monotone-bias assumption belongs to correctness risk, not circularity.

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

The central method rests on standard Monte Carlo facts, two scope assumptions, and one empirically motivated behavioral assumption about estimator bias. No invented entities. The algorithm hyperparameters are hand-chosen defaults rather than fitted constants, but they were selected using the same four benchmarks on which performance is reported, which is the main ledger-side caveat.

free parameters (5)
  • N_epsilon (Rao-Blackwell noise samples per model run) = 100
    Set empirically in §2.2: 'we observed empirically that larger values of N_epsilon typically yield only marginal improvements.' Hand-chosen, fixed across examples, and not covered by the Appendix C sensitivity study.
  • q (elimination probability threshold) = 0.99
    Default selected in §2.4; Appendix C varies q in {0.95, 0.99, 0.999} on the same four test problems used for the main performance claims.
  • N_C (consecutive elimination iterations) = 3
    Default selected in §2.4; Appendix C compares N_C=2 vs N_C=3 on the test suite.
  • N_new (new samples per iteration) = 8
    Default selected in §2.4; Appendix C compares N_new=8 vs N_new=16 on the test suite.
  • N_B (bootstrap replicates) = 100000
    Fixed default in §2.4; large enough that bootstrap Monte Carlo error is negligible relative to the probability threshold q.
assumptions (4)
  • ad hoc to paper The bias of NMC/LRNMC/ANMC is positive and monotonically increasing in the true EIG, so it 'amplifies differences between competing designs' (§2.3).
    Demonstrated empirically on one lognormal test problem (Figure 1) and invoked as the reason ANMC-based elimination is reliable and RNMC-based is not. This is load-bearing for Algorithm 1's elimination decisions: if the monotonicity fails, bootstrap confidence in (16) is confidence in a possibly wrong ordering.
  • domain assumption Likelihood evaluations p(y|θ,ξ) are analytically available/cheap given a model output, and noise sampling costs are negligible relative to model evaluations.
    Equations (7)–(12) require evaluating log p(y_i|θ_j,ξ) for arbitrary prior samples, and ANMC further averages over N_epsilon=100 noise draws per outer sample. This excludes implicit-likelihood simulators where the density is not available. The assumption is stated in §2 but restricts the method's domain.
  • standard math Standard Monte Carlo consistency, NMC bias results, and bootstrap consistency for the paired differences in (15).
    Bias and MSE-rate claims are inherited from [26]; the bootstrap estimate (16) presumes resampling consistently estimates the distribution of the (biased) estimator difference. These are standard tools, not new content.
  • domain assumption The elimination rule (bootstrap probability above q for NC consecutive iterations) is a valid criterion for discarding the true best design only if the estimator ordering matches the true ordering at the sample sizes used.
    Algorithm 1 treats a high bootstrap probability that the current reference beats a challenger as evidence the challenger is truly inferior. This presumes the biased estimator's ranking is the true ranking, which is exactly the property the paper validates only empirically.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Efficient Bayesian Optimal Experimental Design for Expensive Computational Models over Finite Design Sets." pith.science (2026). https://pith.science/paper/HHX4PGYK

@misc{pith2026260716933,
  author       = {Pith},
  title        = {Pith review of: Efficient Bayesian Optimal Experimental Design for Expensive Computational Models over Finite Design Sets},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HHX4PGYK}},
  note         = {Machine review of arXiv:2607.16933}
}
read the original abstract

Bayesian calibration is a powerful framework for identifying parameters in complex and large-scale computational models. However, when there is insufficient or poorly suited data for the calibration process, significant uncertainty about the identified parameters remains. This uncertainty can hinder effective decision-making and understanding of the system. In many applications, the experiment that generates the data can be influenced by design variables, such as sensor placements, loading conditions, or test configurations. Often, however, these design variables can not be chosen arbitrarily but only from a finite set of possible experimental designs. We propose an adaptive algorithm for Bayesian optimal experimental design over a finite design set, specifically tailored for applications involving expensive computational models. The method integrates an accelerated nested Monte Carlo estimator that reuses parameter samples to reduce model evaluations. Additionally, it employs common random numbers and Rao--Blackwellization to reduce variance in pairwise expected information gain comparisons. To support decision-making within the algorithm, we estimate the probability that one design outperforms another using bootstrap sampling. Starting with small sample sizes, the algorithm iteratively eliminates inferior designs based on these probabilistic comparisons, allocating additional computational effort only to promising candidates until a single design remains. The resulting approach achieves high reliability at substantially reduced computational cost, making it well-suited for large-scale engineering applications.

Figures

Figures reproduced from arXiv: 2607.16933 by the authors.

Figure 1
Figure 1. Bias of the RNMC and ANMC estimator of the EIG for a problem with lognormal prior and [PITH_FULL_IMAGE:figures/full_fig_p006_1.png] view at source ↗
Figure 2
Figure 2. SDC (with and without CRN) and ADE performance over 1000 independent runs with the different [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. Sampling schedules for the pharmacokinetic model. For each scheme (geometric, uniform, beta), [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figures from the paper (10 more)
Figure 4
Figure 4. Figure 4: SDC (with and without CRN) and ADE performance over 1000 independent runs with the different [PITH_FULL_IMAGE:figures/full_fig_p010_4.png]
Figure 5
Figure 5. Figure 5: Segmented human lung (left) that is used for the generation of the computational model (right). [PITH_FULL_IMAGE:figures/full_fig_p011_5.png]
Figure 6
Figure 6. Figure 6: Ventilation pressure protocols of the four experimental designs are shown in the top row, and the [PITH_FULL_IMAGE:figures/full_fig_p012_6.png]
Figure 7
Figure 7. Figure 7: SDC (with and without CRN) and ADE performance over 1000 runs with the different estimators [PITH_FULL_IMAGE:figures/full_fig_p012_7.png]
Figure 8
Figure 8. Figure 8: Geometry and finite element mesh of the specimen. The purple arrows indicate the different [PITH_FULL_IMAGE:figures/full_fig_p014_8.png]
Figure 9
Figure 9. Figure 9: SDC (with CRN) and ADE performance over 1000 runs with the different estimators for the [PITH_FULL_IMAGE:figures/full_fig_p015_9.png]
Figure 10
Figure 10. Figure 10: ADE performance over 1000 independent runs with the different estimators and with varying [PITH_FULL_IMAGE:figures/full_fig_p021_10.png]
Figure 11
Figure 11. Figure 11: ADE performance over 1000 independent runs with the different estimators and with varying [PITH_FULL_IMAGE:figures/full_fig_p022_11.png]
Figure 12
Figure 12. Figure 12: ADE performance over 1000 runs with the different estimators and with varying hyperparameter [PITH_FULL_IMAGE:figures/full_fig_p023_12.png]
Figure 13
Figure 13. Figure 13: ADE performance over 1000 runs with the different estimators and with varying hyperparameter [PITH_FULL_IMAGE:figures/full_fig_p024_13.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

56 extracted references · 10 canonical work pages

  1. [1]

    Bayesian Calibration of Computer Models

    M. C. Kennedy and A. O’Hagan. “Bayesian Calibration of Computer Models”. In:Journal of the Royal Statistical Society Series B: Statistical Methodology63.3 (Sept. 2001), pp. 425–464.doi:10.1111/1467- 9868.00294

    arXiv 2001

  2. [2]

    Multi-physics-enhanced Bayesian inverse analysis: Information gain from additional fields

    L. J. Haeusel, J. Nitzler, L. J. Köglmeier, and W. A. Wall. “Multi-physics-enhanced Bayesian inverse analysis: Information gain from additional fields”. In:Computer Methods in Applied Mechanics and Engineering452 (Apr. 2026), p. 118735.doi:10.1016/j.cma.2026.118735

    arXiv 2026

  3. [3]

    On a Measure of the Information Provided by an Experiment

    D. V. Lindley. “On a Measure of the Information Provided by an Experiment”. In:The Annals of Mathematical Statistics27.4 (Dec. 1956), pp. 986–1005.doi:10.1214/aoms/1177728069

    arXiv 1956

  4. [4]

    D. V. Lindley.Bayesian Statistics: A Review. Society for Industrial and Applied Mathematics, Jan. 1972.doi:10.1137/1.9781611970654

    doi:10.1137/1.9781611970654 1972

  5. [5]

    Expected Information as Expected Utility

    J. M. Bernardo. “Expected Information as Expected Utility”. In:The Annals of Statistics7.3 (May 1979).doi:10.1214/aos/1176344689. 16

    arXiv 1979

  6. [6]

    Bayesian Experimental Design: A Review

    K. Chaloner and I. Verdinelli. “Bayesian Experimental Design: A Review”. In:Statistical Science10.3 (Aug. 1995).doi:10.1214/ss/1177009939

    arXiv 1995

  7. [7]

    Maximum Entropy Sampling and Optimal Bayesian Experimental Design

    P. Sebastiani and H. P. Wynn. “Maximum Entropy Sampling and Optimal Bayesian Experimental Design”. In:Journal of the Royal Statistical Society Series B: Statistical Methodology62.1 (Jan. 2000), pp. 145–157.doi:10.1111/1467-9868.00225

    arXiv

  8. [8]

    A Review of Modern Computational Algorithms for Bayesian Optimal Design

    E. G. Ryan, C. C. Drovandi, J. M. McGree, and A. N. Pettitt. “A Review of Modern Computational Algorithms for Bayesian Optimal Design”. In:International Statistical Review84.1 (Apr. 2016), pp. 128– 154.doi:10.1111/insr.12107

    doi:10.1111/insr.12107 2016

Show all 56 references
  1. [9]

    Modern Bayesian Experimental Design

    T. Rainforth, A. Foster, D. R. Ivanova, and F. Bickford Smith. “Modern Bayesian Experimental Design”. In:Statistical Science39.1 (Feb. 2024).doi:10.1214/23-STS915

    2024 doi

  2. [10]

    Optimal experimental design: Formulations and computations

    X. Huan, J. Jagalur, and Y. Marzouk. “Optimal experimental design: Formulations and computations”. In:Acta Numerica33 (July 2024), pp. 715–840.doi:10.1017/S0962492924000023

    2024 doi

  3. [11]

    A. C. Atkinson, A. N. Donev, and R. D. Tobias.Optimum Experimental Designs, with SAS. Oxford statistical science series 34. Oxford: Oxford University Press, 2007

    2007

  4. [12]

    Optimal Designs for Nonlinear Models With Correlated Errors

    J. M. Hughes-Olivier. “Optimal Designs for Nonlinear Models With Correlated Errors”. In:Institute of Mathematical Statistics Lecture Notes - Monograph Series. Hayward, CA: Institute of Mathematical Statistics, 1998, pp. 163–174.doi:10.1214/lnms/1215456195

    1998

  5. [13]

    Minimax optimal designs via particle swarm optimization methods

    R.-B. Chen, S.-P. Chang, W. Wang, H.-C. Tung, and W. K. Wong. “Minimax optimal designs via particle swarm optimization methods”. In:Statistics and Computing25.5 (Sept. 2015), pp. 975–988. doi:10.1007/s11222-014-9466-0

    2015 doi

  6. [14]

    Optimal Bayesian design applied to logistic regression experiments

    K. Chaloner and K. Larntz. “Optimal Bayesian design applied to logistic regression experiments”. In: Journal of Statistical Planning and Inference21.2 (Feb. 1989), pp. 191–208.doi: 10.1016/0378- 3758(89)90004-9

    1989 doi

  7. [15]

    Parameter Neutral Optimum Design for Non-linear Models

    D. Firth and J. P. Hinde. “Parameter Neutral Optimum Design for Non-linear Models”. In:Journal of the Royal Statistical Society: Series B (Methodological)59.4 (Nov. 1997), pp. 799–811.doi:10.1111/1467- 9868.00097

    1997

  8. [16]

    Sequential Optimal Design of Neurophysiology Experiments

    J. Lewi, R. Butera, and L. Paninski. “Sequential Optimal Design of Neurophysiology Experiments”. In: Neural Computation21.3 (Mar. 2009), pp. 619–687.doi:10.1162/neco.2008.08-07-594

    2009 doi

  9. [17]

    Fast estimation of expected information gains for Bayesian experimental designs based on Laplace approximations

    Q. Long, M. Scavino, R. Tempone, and S. Wang. “Fast estimation of expected information gains for Bayesian experimental designs based on Laplace approximations”. In:Computer Methods in Applied Mechanics and Engineering259 (June 2013), pp. 24–39.doi:10.1016/j.cma.2013.02.017

    2013 doi

  10. [18]

    Fast Bayesian optimal experimental design for seismic source inversion

    Q. Long, M. Motamed, and R. Tempone. “Fast Bayesian optimal experimental design for seismic source inversion”. In:Computer Methods in Applied Mechanics and Engineering291 (July 2015), pp. 123–145. doi:10.1016/j.cma.2015.03.021

    2015 doi

  11. [19]

    A Laplace method for under-determined Bayesian optimal experimental designs

    Q. Long, M. Scavino, R. Tempone, and S. Wang. “A Laplace method for under-determined Bayesian optimal experimental designs”. In:Computer Methods in Applied Mechanics and Engineering285 (Mar. 2015), pp. 849–876.doi:10.1016/j.cma.2014.12.008

    2015 doi

  12. [20]

    Multimodal information gain in Bayesian design of experiments

    Q. Long. “Multimodal information gain in Bayesian design of experiments”. In:Computational Statistics 37.2 (Apr. 2022), pp. 865–885.doi:10.1007/s00180-021-01145-9

    2022 doi

  13. [21]

    Fully Bayesian Experimental Design for Pharmacokinetic Studies

    E. Ryan, C. Drovandi, and A. Pettitt. “Fully Bayesian Experimental Design for Pharmacokinetic Studies”. In:Entropy17.3 (Mar. 2015), pp. 1063–1089.doi:10.3390/e17031063

    2015 doi

  14. [22]

    Fast Bayesian experimental design: Laplace-based importance sampling for the expected information gain

    J. Beck, B. M. Dia, L. F. Espath, Q. Long, and R. Tempone. “Fast Bayesian experimental design: Laplace-based importance sampling for the expected information gain”. In:Computer Methods in Applied Mechanics and Engineering334 (June 2018), pp. 523–553.doi:10.1016/j.cma.2018.01.053

    2018 doi

  15. [23]

    Approximating Large-Scale Hessian Matrices Using Secant Equations

    J. M. Fowkes, N. I. M. Gould, and J. A. Scott. “Approximating Large-Scale Hessian Matrices Using Secant Equations”. In:ACM Transactions on Mathematical Software51.2 (June 2025), pp. 1–21.doi: 10.1145/3728460

    2025 doi

  16. [24]

    Towards Bayesian experimental design for nonlinear models that require a large number of sampling times

    E. G. Ryan, C. C. Drovandi, M. H. Thompson, and A. N. Pettitt. “Towards Bayesian experimental design for nonlinear models that require a large number of sampling times”. In:Computational Statistics & Data Analysis70 (Feb. 2014), pp. 45–60.doi:10.1016/j.csda.2013.08.017

    2014 doi

  17. [25]

    Estimating Expected Information Gains for Experimental Designs With Application to the Random Fatigue-Limit Model

    K. J. Ryan. “Estimating Expected Information Gains for Experimental Designs With Application to the Random Fatigue-Limit Model”. In:Journal of Computational and Graphical Statistics12.3 (Sept. 2003), pp. 585–603.doi:10.1198/1061860032012

    2003 doi

  18. [26]

    On Nesting Monte Carlo Estimators

    T. Rainforth, R. Cornish, H. Yang, A. Warrington, and F. Wood. “On Nesting Monte Carlo Estimators”. In:Proceedings of the 35th International Conference on Machine Learning. Ed. by J. Dy and A. Krause. Vol. 80. Proceedings of Machine Learning Research. PMLR, July 2018, pp. 4267...

    2018

  19. [27]

    Variational Bayesian optimal experimental design

    A. Foster, M. Jankowiak, E. Bingham, P. Horsfall, Y. W. Teh, T. Rainforth, and N. Goodman. “Variational Bayesian optimal experimental design”. In:Proceedings of the 33rd International Conference on Neural Information Processing Systems. Red Hook, NY, USA: Curran Associates Inc., 2019

    2019

  20. [28]

    Multilevel Monte Carlo Path Simulation

    M. B. Giles. “Multilevel Monte Carlo Path Simulation”. In:Operations Research56.3 (June 2008), pp. 607–617.doi:10.1287/opre.1070.0496

    2008

  21. [29]

    Unbiased MLMC Stochastic Gradient-Based Optimization of Bayesian Experimental Designs

    T. Goda, T. Hironaka, W. Kitade, and A. Foster. “Unbiased MLMC Stochastic Gradient-Based Optimization of Bayesian Experimental Designs”. In:SIAM Journal on Scientific Computing44.1 (Feb. 2022), A286–A311.doi:10.1137/20M1338848

    2022 doi

  22. [30]

    Unbiased Estimation with Square Root Convergence for SDE Models

    C.-H. Rhee and P. W. Glynn. “Unbiased Estimation with Square Root Convergence for SDE Models”. In:Operations Research63.5 (Oct. 2015), pp. 1026–1043.doi:10.1287/opre.2015.1404

    2015

  23. [31]

    Variational Bayesian optimal experimental design with normalizing flows

    J. Dong, C. Jacobsen, M. Khalloufi, M. Akram, W. Liu, K. Duraisamy, and X. Huan. “Variational Bayesian optimal experimental design with normalizing flows”. In:Computer Methods in Applied Mechanics and Engineering433 (Jan. 2025), p. 117457.doi:10.1016/j.cma.2024.117457

    2025

  24. [32]

    Practical Bayesian Optimization of Machine Learning Al- gorithms

    J. Snoek, H. Larochelle, and R. Adams. “Practical Bayesian Optimization of Machine Learning Al- gorithms”. In:Advances in Neural Information Processing Systems. Ed. by F. Pereira, C. J. Burges, L. Bottou, and K. Weinberger. Vol. 25. Curran Associates, Inc., 2012

    2012

  25. [33]

    OnEstimatingtheGradientoftheExpectedInformationGaininBayesianExperimental Design

    Z.AoandJ.Li.“OnEstimatingtheGradientoftheExpectedInformationGaininBayesianExperimental Design”. In:Proceedings of the AAAI Conference on Artificial Intelligence38.18 (Mar. 2024), pp. 20311– 20319.doi:10.1609/aaai.v38i18.30012

    2024 doi

  26. [34]

    Accelerated Bayesian experimental design for chemical kinetic models

    X. Huan. “Accelerated Bayesian experimental design for chemical kinetic models”. Master’s thesis. Massachusetts Institute of Technology, 2010

    2010

  27. [35]

    Simulation-based optimal Bayesian experimental design for nonlinear systems

    X. Huan and Y. M. Marzouk. “Simulation-based optimal Bayesian experimental design for nonlinear systems”. In:Journal of Computational Physics232.1 (Jan. 2013), pp. 288–317.doi: 10.1016/j.jcp. 2012.08.013

    2013 doi

  28. [36]

    Some Guidelines and Guarantees for Common Random Numbers

    P. Glasserman and D. D. Yao. “Some Guidelines and Guarantees for Common Random Numbers”. In: Management Science38.6 (June 1992), pp. 884–908.doi:10.1287/mnsc.38.6.884

    1992 doi

  29. [37]

    R. Y. Rubinstein and D. P. Kroese.Simulation and the Monte Carlo Method. 1st ed. Wiley Series in Probability and Statistics. Wiley, Nov. 2016.doi:10.1002/9781118631980

    2016 doi

  30. [38]

    C. P. Robert and G. Casella.Monte Carlo Statistical Methods. Springer Texts in Statistics. New York, NY: Springer New York, 2004.doi:10.1007/978-1-4757-4145-2

    2004 doi

  31. [39]

    Efron and R

    B. Efron and R. Tibshirani.An Introduction to the Bootstrap. 0th ed. Chapman and Hall/CRC, May 1994.doi:10.1201/9780429246593

    1994 doi

  32. [40]

    A Mathematical Theory of Communication

    C. E. Shannon. “A Mathematical Theory of Communication”. In:Bell System Technical Journal27.3 (July 1948), pp. 379–423.doi:10.1002/j.1538-7305.1948.tb01338.x

    1948

  33. [41]

    A Scalable Gradient Free Method for Bayesian Experimental Design with Implicit Models

    J. Zhang, S. Bi, and G. Zhang. “A Scalable Gradient Free Method for Bayesian Experimental Design with Implicit Models”. In:Proceedings of The 24th International Conference on Artificial Intelligence and Statistics. Ed. by A. Banerjee and K. Fukumizu. Vol. 130. Proceedings of M...

    2021

  34. [42]

    Coupled and reduced dimensional modeling of respiratory mechanicsduringspontaneousbreathing

    M. Ismail, A. Comerford, and W. A. Wall. “Coupled and reduced dimensional modeling of respiratory mechanicsduringspontaneousbreathing”.In:International Journal for Numerical Methods in Biomedical Engineering29.11 (2013), pp. 1285–1305.doi:10.1002/cnm.2577

    2013 doi

  35. [43]

    A comprehensive computational human lung model incorporating inter-acinar dependencies: Application to spontaneous breathing and mechanical ventilation

    C. J. Roth, M. Ismail, L. Yoshihara, and W. A. Wall. “A comprehensive computational human lung model incorporating inter-acinar dependencies: Application to spontaneous breathing and mechanical ventilation”. In:International Journal for Numerical Methods in Biomedical Engineer...

    2017 doi

  36. [44]

    Pressure- and time-dependent alveolar recruitment/derecruitment in a spatially resolved patient-specific computational model for injured human lungs

    C. M. Geitner, L. J. Köglmeier, I. Frerichs, P. Langguth, M. Lindner, D. Schädler, N. Weiler, T. Becher, and W. A. Wall. “Pressure- and time-dependent alveolar recruitment/derecruitment in a spatially resolved patient-specific computational model for injured human lungs”. In:I...

    2024 doi

  37. [45]

    Patient-specific prediction of regional lung mechanics in patients with ARDS with physics-based models: a validation study

    M. Rixner, M. Ludwig, M. Lindner, I. Frerichs, A. Sablewski, K.-R. Wichmann, M.-C. Wachter, K. W. Müller, D. Schädler, W. A. Wall, J. Biehler, and T. Becher. “Patient-specific prediction of regional lung mechanics in patients with ARDS with physics-based models: a validation s...

    2025

  38. [46]

    Sablewski, M

    A. Sablewski, M. Ludwig, C. Eichinger, M. Lindner, P. Langguth, D. Schädler, W. A. Wall, and T. Becher. “Predicting Global and Regional Respiratory Response in ARDS Patients through Individualized Physics- Based Computational Lung Models: A Prospective Clinical Pilot Study [Ma...

  39. [47]

    Adaptive Estimate Interpolation: Accelerating Local Newton-Raphson Schemes in Computational Plasticity and Viscoplasticity [Manuscript in preparation]

    D. C. Ana, C. P. Schmidt, and W. A. Wall. “Adaptive Estimate Interpolation: Accelerating Local Newton-Raphson Schemes in Computational Plasticity and Viscoplasticity [Manuscript in preparation]”

  40. [48]

    A thermodynamically consistent formulation of the Johnson–Cook model

    C. Mareau. “A thermodynamically consistent formulation of the Johnson–Cook model”. In:Mechanics of Materials143 (Apr. 2020), p. 103340.doi:10.1016/j.mechmat.2020.103340

    2020

  41. [49]

    4C.4C: A Comprehensive Multiphysics Simulation Framework.url: https://www.4c-multiphysics. org

  42. [50]

    Biehler, J

    J. Biehler, J. Nitzler, S. Brandstaeter, M. Dinkel, V. Gravemeier, L. J. Haeusel, G. R. Rei, H. Willmann, B. Wirthl, and W. A. Wall.QUEENS: An Open-Source Python Framework for Solver-Independent Analyses of Large-Scale Computational Models. 2025.url:https://arxiv.org/abs/2508.16316

    2025 arXiv

  43. [51]

    Goal-Oriented Bayesian Optimal Experimental Design for Nonlinear Models Using Markov Chain Monte Carlo

    S. Zhong, W. Shen, T. Catanach, and X. Huan. “Goal-Oriented Bayesian Optimal Experimental Design for Nonlinear Models Using Markov Chain Monte Carlo”. In:SIAM/ASA Journal on Uncertainty Quantification14.1 (2026), pp. 19–47.doi:10.1137/24M1649344

    2026 doi

  44. [52]

    Small-noise approximation for Bayesian optimal experimental design with nuisance uncertainty

    A. Bartuska, L. Espath, and R. Tempone. “Small-noise approximation for Bayesian optimal experimental design with nuisance uncertainty”. In:Computer Methods in Applied Mechanics and Engineering399 (Sept. 2022), p. 115320.doi:10.1016/j.cma.2022.115320

    2022

  45. [53]

    Feng and Y

    C. Feng and Y. M. Marzouk.A layered multiple importance sampling scheme for focused optimal Bayesian experimental design. 2019.url:https://arxiv.org/abs/1903.11187

    2019 arXiv

  46. [54]

    Estimating mutual information

    A. Kraskov, H. Stögbauer, and P. Grassberger. “Estimating mutual information”. In:Physical Review E 69.6 (June 2004), p. 066138.doi:10.1103/PhysRevE.69.066138

    2004 doi

  47. [55]

    C. M. Bishop.Pattern recognition and machine learning. Information science and statistics. New York: Springer, 2006

    2006

  48. [56]

    K. P. Murphy.Machine learning: a probabilistic perspective. MIT press, 2012. A Derivation of Expected Information Gain As shown in [9], the expected information gain can be derived as: EIG(ξ) =E p(y|ξ) IG(ξ,y) =E p(y|ξ) H[p(θ)]−H[p(θ|y,ξ)] = Z p(y|ξ) p(θ|y,ξ) logp(θ|y,ξ)−p(θ) ...

    2012

Pith tools

Reviewed August 1, 2026 · model on record in the stance chip above.