Pith. sign in

REVIEW 2 major objections 6 minor 1 cited by

A Framework for the Bayesian Calibration of Complex and Data-Scarce Models in Applied Sciences

T0 review · 2 major / 6 minor · reviewed 2026-08-03 · deepseek-v4-flash

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

Pith's one-line read The paper claims a single extensible Python library, ACBICI, can unify all KOH calibration variants—surrogate, discrepancy, single- and multi-output—for data-scarce engineering models.

desk verdict A competent KOH review wrapped around a potentially useful Python package, but the multi-output kernel story is internally inconsistent and all validation is synthetic. read the letter →

arxiv 2601.22890 v1 pith:EWWYYD3D submitted 2026-01-30 stat.CO cond-mat.mtrl-scimath.OCmath.STstat.TH

Christina Schenk , Ignacio Romero This is my paper
classification stat.COcond-mat.mtrl-scimath.OCmath.STstat.TH MSC 62F1565C60
keywords BayesiancalibrationKOHGaussianprocesssurrogatesmodeldiscrepancymulti-outputscarcedatauncertaintyquantificationPythonlibrary
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's central claim is that Bayesian calibration of computer models can be organized into a small set of types (simple, surrogate-based, discrepancy-corrected, and combined) and that all of them—for both single and multiple outputs—can be implemented consistently in one open-source Python library, ACBICI. If that claim holds, engineers and applied scientists with expensive simulations and scarce experimental data can get posterior distributions for their parameters, including structural model discrepancy, without becoming Bayesian specialists. The paper also assembles a practical checklist of scaling, prior-selection, and convergence-diagnostic guidelines that are scattered through the literature, and it illustrates the workflow on gravity, production-function, and traction-test examples. The multi-output extension works by adding a task index to the input space and using a block-diagonal covariance kernel that treats outputs as independent—an assumption the paper itself calls strong.

What carries the argument

The load-bearing object is the KOH calibration likelihood with its Gaussian-process covariance structure, extended to multiple outputs by augmenting the input space with a task index and imposing a block-diagonal kernel that assigns zero covariance across tasks (Eq. 28). This construction preserves the single-output machinery for vector-valued responses, making all four calibration types available for multi-output problems. Supporting components are the GP surrogate and discrepancy kernels, non-dimensionalization and scaling of inputs, the two posterior samplers (ensemble MCMC and variational Bayesian Monte Carlo), and the diagnostic suite (split-R-hat, ESS, autocorrelation, corner plots, pr

What would settle it

Use a two-output synthetic problem generated with a known cross-output correlation (for example, from a coregionalized Gaussian process) and known parameters. Calibrate it with ACBICI's multi-output KOH method and compare against two independent single-output calibrations. If the multi-output posterior is not tighter or more accurate—or if it is overconfident because the zero-covariance assumption is violated—the claimed advantage of the multi-output method is refuted.

Watch

Extended reading notes

Core claim

On its own terms, the paper's contribution is a unified and extensible implementation of KOH Bayesian calibration. The key move is to reduce multi-output calibration to a single-output problem by appending a task index to the input coordinates and using a covariance kernel that is zero for pairs with different task indices. This lets all four calibration types (A–D, with or without estimating experimental error variance) operate on vector-valued outputs. The framework pairs Gaussian-process surrogates and discrepancy terms with two posterior approximations—MCMC and variational Bayesian Monte Carlo—and bundles diagnostic plots and convergence statistics into a short command sequence. The arti

Load-bearing premise

The multi-output extension assumes different outputs are statistically independent, using a block-diagonal kernel with zero cross-task covariance; if outputs actually share information, the joint calibration throws away the shared signal that motivated it.

Editorial extensions

If this is right

  • If ACBICI works as claimed, a calibration that previously required a bespoke Bayesian pipeline becomes a five-command workflow: define the model, load data, choose a calibration type, run, and inspect diagnostics.
  • Expensive simulators can be calibrated with GP surrogates while retaining quantified surrogate uncertainty, so practitioners get posterior intervals instead of point estimates.
  • Multi-output calibration is available in all four calibration types, so coupled physical quantities can be inferred jointly from scarce data—subject to the paper's stated independence assumption.
  • The built-in diagnostics (split-R-hat, effective sample size, autocorrelation time, ELBO stability) give practitioners standard checks for convergence and sampling quality in both MCMC and VBMC modes.
  • The consolidated practical recommendations give new users a concrete path on scaling, prior choices, and when to use discrepancy-aware calibration.

Reading between the lines

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

  • Because the paper's multi-output kernel forces zero covariance between tasks (Eq. 28), a natural test is to replace it with a coregionalization kernel that estimates cross-output correlations; if correlated outputs exist, such a kernel should reduce posterior width or improve parameter recovery compared with the paper's block-diagonal version.
  • The paper's own gravity example demonstrates a practical pitfall: type A calibration can yield a confident but wrong posterior when the model is structurally inadequate, implying that the framework's reliability depends on users choosing discrepancy-aware types when model bias is suspected.
  • The claim of being the first unified single/multi-output KOH library depends on how 'KOH' is defined; since the paper excludes several existing Python tools because they are 'not strictly based on' KOH, the novelty claim is narrower than a reader might assume.
  • The eight-type taxonomy (A–D with or without error-variance estimation) could serve as a benchmark template for comparing calibration software, since each type isolates a specific modeling component.
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

2 major / 6 minor

Summary. The paper reviews the Kennedy–O'Hagan (KOH) framework for Bayesian calibration of computer models and proposes a unified implementation in a new open-source Python library, ACBICI. It defines four calibration types (A–D) differing in whether a surrogate model, a discrepancy term, and/or experimental-error estimation are included; extends the framework to multiple outputs via an augmented input space with a task index; and describes posterior sampling with MCMC (emcee) and variational Bayesian Monte Carlo (PyVBMC). Three examples are presented: a gravity-drop problem (types A and C), a Cobb–Douglas production function (type B), and a traction test with two outputs (type D). The main claimed contributions are a self-contained KOH review, the ACBICI library with single- and multi-output calibration, practical recommendations, and illustrative examples.

Significance. If the implementation is sound, ACBICI could be a useful practical contribution: it is open source, well documented, and provides a unified workflow for standard KOH-style calibration, which is valuable for practitioners in engineering and applied science. The paper's theoretical review of types A–D is correct standard KOH material, and the practical recommendations (scaling, priors, diagnostics, convergence) are sensible and clearly presented. The main novelty is the multi-output capability, but this is also the weakest part: the multi-output kernel in the main text (Eq. (28)) sets cross-task covariance to zero, while the supplementary tracture-test description reports a nonstationary multi-task kernel. The validation is also self-referential — all data are generated with ACBICI and the examples are coded within ACBICI. These issues must be resolved before the central claims can be accepted.

major comments (2)
  1. [§2.8, Eq. (28); SI 'Traction test example'] The multi-output surrogate covariance in Eq. (28) is block-diagonal with zero cross-task covariance for i≠j. This makes the likelihood factorize over tasks and excludes cross-output correlations. However, the SI for the sole multi-output example states that 'a nonstationary multi-task kernel is employed, enabling the Gaussian process surrogate to model both outputs jointly' and reports hyperparameters β_x, β_t, λ_x, β_d, λ_d. These two descriptions are mutually inconsistent. If the implementation uses a nonstationary multi-task kernel, Eq. (28) does not describe the method; if it uses Eq. (28), the SI claim is incorrect and the method cannot exploit correlated outputs — contradicting the introduction's motivation based on 'correlated quantities of interest' (§1). Since the traction test is the only demonstration of multi-output calibration, the central multi-output contribution rests on
  2. [§4; 'Data and code availability'] The validation is self-referential. The data and code availability statement says 'All data used in this study were synthetically generated with ACBICI version 2.1.0', and the examples are implemented as ACBICI model subclasses. Thus the package is used to generate the data, to define the models, and to produce the diagnostics that are then presented as evidence of success. The reported split-Rhat, ESS, and prediction-error statistics show only that the samplers converged on these particular synthetic problems; they do not demonstrate that the calibrated parameters are correct or that the workflow is reliable on independent data. The gravity example even acknowledges that the MAP is a bad estimate (§4.1.1). To substantiate the abstract's claim of 'reliable calibration', the authors should include at least one example with an independent simulator, real or held-out experimental data, or a
minor comments (6)
  1. [Eq. (12) and Eq. (15)] Eq. (12) uses ν without any definition, and Eq. (15) writes π(θ,χ,ν) where σ is intended (compare Eq. (14) and Eq. (20)). These are recognizable typos but should be corrected in a paper meant to be a self-contained reference.
  2. [Eq. (6)] The data set is defined as a subset of R^d × R^p, but y_i is a scalar output. This should be R^d × R (or R^d × R^{n_tasks} for multi-output).
  3. [§4.1, Eq. (38) vs. Eq. (43)] The main text refers to the gravity model as Eq. (43) in §4.1, but the model is Eq. (38) in the main text; Eq. (43) appears only in the SI. Cross-referencing between main text and SI is inconsistent throughout.
  4. [§4.3.2] The text refers to 'Figures 21–23' but in the main text the corresponding figures are numbered 10–12. Figure numbers in the main text and SI are out of sync.
  5. [§4.3, Eq. (42)] The nondimensionalization states β_y2 = y2/L0, but y2 is an area and should be divided by A0 (as done in the SI, Eq. (48)). The main text also inconsistently defines β_F / (E0 A0) with reference to F0.
  6. [Table 1] There is a typo in the table caption: 'T able 1' should be 'Table 1'.

Circularity Check

1 steps flagged · score 4.0 of 10

The multi-output 'extension' in Eq. (28) is equivalent by construction to independent single-output KOH; the SI's nonstationary multi-task kernel contradicts Eq. (28), but the core KOH theory is externally grounded.

  1. self definitional [Section 2.8, Eq. (28); cf. Conclusions (Section 5)]
    "ˆcS(ˆx, ˆx′) = ( cS(x, x′) if i=j , 0 otherwise, (28) ... This block-diagonal construction results in a kernel that is active only within each task, treating outputs as independent and excluding cross-output correlations."

    With ěit c_S = 0 for i≠j, the covariance matrix in the joint likelihood is block-diagonal, so the posterior factorizes as π(θ)∏_i L_i(θ). Thus the 'multi-output' calibration is, by construction, equivalent to independent single-output KOH calibrations combined through a shared prior; no cross-output correlation is learned or used. The Conclusions' claim of 'simultaneous Bayesian inference across multiple, interrelated model outputs' is therefore not a derived result but a restatement of the independence assumption in Eq. (28).

full rationale

The core KOH derivation is standard and externally grounded: the likelihoods and posteriors in Eqs. (7)–(25) follow Kennedy & O'Hagan and standard Gaussian-process references, with no load-bearing self-citation chain. The paper's self-references [34,35] are to its own package repository/documentation and are normal software citations. The one construction that reduces to its own input is the multi-output covariance of §2.8: Eq. (28) sets the cross-task covariance to zero, so the joint likelihood factorizes per task and the advertised 'interrelated' multi-output inference is mathematically the same as independent single-output inference with shared priors. The paper itself concedes 'treating outputs as independent and excluding cross-output correlations,' making the multi-output feature a renaming of single-output analysis rather than a derived multi-output method. Additionally, the SI's traction example reports that 'a nonstationary multi-task kernel is employed, enabling the Gaussian process surrogate to model both outputs jointly,' which directly contradicts Eq. (28); the flagship multi-output demonstration therefore validates a different kernel than the one derived in the paper. This contradiction is an internal-consistency problem rather than a circular reduction. The examples use data synthetically generated with ACBICI, but the gravity example explicitly generates data from a different model, and no theoretical derivation is claimed from the examples, so the self-referential validation is a weakness in evidence rather than a circularity of the derivation. Overall: partial circularity in the multi-output claim (score 4), while the central KOH framework and software implementation retain independent content.

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

All theoretical content is inherited from KOH and Gaussian-process literature; the paper's specific contributions are software and guidance. The ledger lists the domain assumptions the tool silently makes: additive noise/discrepancy, GP representation of both emulator and discrepancy, dimensionless isotropic kernels, and independent tasks in multi-output mode.

free parameters (5)
  • Reference scales E0, L0, A0, F0 (traction example) = 250 MPa, 10 mm, 1 mm^2, 250 N
    Chosen by hand in §4.3 and the SI to non-dimensionalize outputs; the reported posteriors for E and ν are relative to these scales and the calibration is only valid in the dimensionless coordinates the library requires.
  • Reference scales L0, K0 (Cobb–Douglas example) = 50, 200 monetary units
    Chosen 'from the available experimental data' (§4.2) to scale labor and capital; since the library assumes dimensionless inputs, the elasticities α and γ are estimated under this normalization.
  • Prior bounds (all examples) = g~U[7,12], α~U[0.3,1.6], γ~N(0.35,0.1), E~U[0.8,1.6] (non-dim), ν~U[0,0.5]
    Hand-specified priors in the example model classes (§4). Prior choices influence the reported posterior means/credible intervals; the paper gives no principled justification.
  • GP kernel hyperparameters βx, βt, λx, βd, λd, and σ = e.g., βx≈1.24, βt≈1.57, λx≈0.60, σ≈0.057 (Cobb–Douglas); E≈0.998, ν≈0.24, split-R̂<1.012 (traction)
    Estimated during calibration (§4 and SI). They determine the surrogate and discrepancy behavior and therefore the reported calibration results, but they are outputs, not inputs, of the method.
  • ACBICI built-in defaults (kernels, priors, sampler settings) = not specified in manuscript
    The library ships 'built-in default settings that facilitate robust starting points' (§1, §3). Since the values are not disclosed, the user cannot know what assumptions are being silently imposed, and reproducibility of the 'default' workflow is compromised.
assumptions (5)
  • domain assumption The computer model output, experimental data, and discrepancy are related by additive Gaussian structure y = m(x;θ) + δ(x) + ε with ε ~ N(0, σ²).
    Definition of the KOH calibration model in Eqs. (5), (18), (21); the entire framework assumes this additive structure.
  • domain assumption Model discrepancy δ can be represented as a zero-mean Gaussian process.
    §2.5–2.6; if the true discrepancy is non-smooth or non-stationary, the GP assumption biases parameter posteriors (a known KOH criticism, acknowledged indirectly in §2.8's limitation paragraph).
  • domain assumption The emulator (surrogate) is a Gaussian process whose posterior is a valid stand-in for the expensive model.
    §2.3; the likelihood (12) treats the GP as the model without accounting for emulator uncertainty in a fully Bayesian two-level way (it is integrated via hyperparameters χ, but the KOH framework's treatment of emulator uncertainty is itself an assumption).
  • ad hoc to paper All variables are dimensionless and isotropic kernels are adequate.
    §2.4 Eq. (16): 'in ACBICI, all variables are assumed to be dimensionless, drastically simplifying the kernel construction, yet introducing limitations.' The paper itself flags loss of interpretability and sensitivity to arbitrary rescaling.
  • ad hoc to paper Cross-task covariance is zero in multi-output calibration.
    Eq. (28): the kernel vanishes for pairs with different task indices. This is the paper's design choice, admitted to be 'a strong assumption' and contradicted by the SI's description of a nonstationary multi-task kernel.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Framework for the Bayesian Calibration of Complex and Data-Scarce Models in Applied Sciences." pith.science (2026). https://pith.science/paper/EWWYYD3D

@misc{pith2026260122890,
  author       = {Pith},
  title        = {Pith review of: A Framework for the Bayesian Calibration of Complex and Data-Scarce Models in Applied Sciences},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/EWWYYD3D}},
  note         = {Machine review of arXiv:2601.22890}
}
read the original abstract

In this work, we review the theory involved in the Bayesian calibration of complex computer models, with particular emphasis on their use for applications involving computationally expensive simulations and scarce experimental data. In the article, we present a unified framework that incorporates various Bayesian calibration methods, including well-established approaches. Furthermore, we describe their implementation and use with a new, open-source Python library, ACBICI (A Configurable BayesIan Calibration and Inference Package). All algorithms are implemented with an object-oriented structure designed to be both easy to use and readily extensible. In particular, single-output and multiple-output calibration are addressed in a consistent manner. The article completes the theory and its implementation with practical recommendations for calibrating the problems of interest. These guidelines -- currently unavailable in a unified form elsewhere -- together with the open-source Python library, are intended to support the reliable calibration of computational codes and models commonly used in engineering and related fields. Overall, this work aims to serve both as a comprehensive review of the statistical foundations and (computational) tools required to perform such calculations, and as a practical guide to Bayesian calibration with modern software tools.

Discussion (0). Sign in to comment.

Forward citations

Cited by 1 Pith paper

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

  1. Bayesian reversal of the liquid level trajectory in a draining tank for pollution forensics

    stat.AP 2026-05 unverdicted novelty 5.0 of 10

    Bayesian inversion infers initial liquid level in a draining tank from final level and duration via Torricelli's law augmented by an empirical discrepancy function, validated on water tank experiments.

Reference graph

Works this paper leans on

76 extracted references · 11 canonical work pages · cited by 1 Pith paper

  1. [1]

    WIREs Computational Statistics16(1), 1645 (2024) https://doi.org/10.1002/wics.1645 https://wires.onlinelibrary.wiley.com/doi/pdf/10.1002/wics.1645

    Sung, C.-L., Tuo, R.: A review on computer model calibration. WIREs Computational Statistics16(1), 1645 (2024) https://doi.org/10.1002/wics.1645 https://wires.onlinelibrary.wiley.com/doi/pdf/10.1002/wics.1645

    doi:10.1002/wics.1645 2024

  2. [2]

    Journal of the Royal Statistical Society: Series B (Statistical Methodology)63(3), 425–464 (2001)

    Kennedy, M.C., O’Hagan, A.: Bayesian calibration of computer models. Journal of the Royal Statistical Society: Series B (Statistical Methodology)63(3), 425–464 (2001)

    2001

  3. [3]

    Journal of Computational Physics545, 114469 (2026) https://doi.org/10.1016/ j.jcp.2025.114469

    Yang, H., Dong, X., Wu, J.-L.: Bayesian experimental design for model discrep- ancy calibration: An auto-differentiable ensemble Kalman inversion approach. Journal of Computational Physics545, 114469 (2026) https://doi.org/10.1016/ j.jcp.2025.114469 . Accessed 2025-12-16

    arXiv 2026

  4. [4]

    https://arxiv.org/abs/2509.18998

    Schenk, C., Jim´ enez, J.A., Romero, I.: Bayesian Calibration and Model Assess- ment of Cell Migration Dynamics with Surrogate Model Integration (2025). https://arxiv.org/abs/2509.18998

    arXiv 2025

  5. [5]

    Automatica39(4), 669–676 (2003) https://doi.org/10

    De Nicolao, G., Ferrari-Trecate, G.: Regularization networks for inverse problems: A state-space approach. Automatica39(4), 669–676 (2003) https://doi.org/10. 1016/S0005-1098(02)00280-7

    2003

  6. [6]

    PLOS Computational Biology19(11), 1–26 (2023) https: //doi.org/10.1371/journal.pcbi.1011111

    Backman, T.W.H., Schenk, C., Radivojevic, T., Ando, D., Singh, J., Czajka, J.J., Costello, Z., Keasling, J.D., Tang, Y., Akhmatskaya, E., Garcia Martin, H.: 34 Bayflux: A Bayesian method to quantify metabolic fluxes and their uncertainty at the genome scale. PLOS Computational Biology19(11), 1–26 (2023) https: //doi.org/10.1371/journal.pcbi.1011111

    work page doi:10.1371/journal.pcbi.1011111 2023

  7. [7]

    SIAM Journal on Scientific Computing26(2), 448–466 (2004) https://doi.org/10.1137/ S1064827503426693

    Higdon, D., Kennedy, M., Cavendish, J.C., Cafeo, J.A., Ryne, R.D.: Combin- ing Field Data and Computer Simulations for Calibration and Prediction. SIAM Journal on Scientific Computing26(2), 448–466 (2004) https://doi.org/10.1137/ S1064827503426693

    2004

  8. [8]

    Journal of the American Statistical Association 103(482), 570–583 (2008) https://doi.org/10.1198/016214507000000888

    Higdon, D., Gattiker, J., Williams, B., Rightley, M.: Computer model calibration using high-dimensional output. Journal of the American Statistical Association 103(482), 570–583 (2008) https://doi.org/10.1198/016214507000000888

    doi:10.1198/016214507000000888 2008

Show all 76 references
  1. [9]

    Archives of Computational Methods in Engineering27(2), 361–385 (2020) https://doi.org/ 10.1007/s11831-018-09311-x

    Rappel, H., Beex, L.A.A., Hale, J.S., Noels, L., Bordas, S.P.A.: A Tutorial on Bayesian Inference to Identify Material Parameters in Solid Mechanics. Archives of Computational Methods in Engineering27(2), 361–385 (2020) https://doi.org/ 10.1007/s11831-018-09311-x

    2020 doi

  2. [10]

    Journal of the Mechanics and Physics of Solids149, 104284 (2021) https://doi.org/10.1016/j.jmps.2020

    Nguyen, T., Francom, D.C., Luscher, D.J., Wilkerson, J.W.: Bayesian calibration of a physics-based crystal plasticity and damage model. Journal of the Mechanics and Physics of Solids149, 104284 (2021) https://doi.org/10.1016/j.jmps.2020. 104284

    2021 doi

  3. [11]

    Bayesian Analysis1(4), 765–792 (2006) https://doi

    Gattiker, J., Higdon, D., Keller-McNulty, S., McKay, M., Moore, L., Williams, B.: Combining experimental data and computer simulations, with an application to flyer plate experiments. Bayesian Analysis1(4), 765–792 (2006) https://doi. org/10.1214/06-BA125

    2006 doi

  4. [12]

    Reliability Engineering & System Safety91(10), 1290–1300 (2006) https://doi.org/10.1016/ j.ress.2005.11.025

    O’Hagan, A.: Bayesian analysis of computer code outputs: A tutorial. Reliability Engineering & System Safety91(10), 1290–1300 (2006) https://doi.org/10.1016/ j.ress.2005.11.025 . The Fourth International Conference on Sensitivity Analysis of Model Output (SAMO 2004)

    2006

  5. [13]

    SAE International Journal of Passenger Cars - Mechanical Systems8(2), 415–420 (2015) https://doi.org/10

    Bergerson, J., Muehleisen, R.: Bayesian Large Model Calibration Using Simula- tion and Measured Data for Improved Predictions. SAE International Journal of Passenger Cars - Mechanical Systems8(2), 415–420 (2015) https://doi.org/10. 4271/2015-01-0481 . Publisher: SAE Internatio...

    2015

  6. [14]

    Energy and Buildings174, 527–547 (2018) https://doi.org/10

    Chong, A., Menberg, K.: Guidelines for the Bayesian calibration of building energy models. Energy and Buildings174, 527–547 (2018) https://doi.org/10. 1016/j.enbuild.2018.06.028

    2018

  7. [15]

    Riddle, M., Muehleisen, R.: A guide to Bayesian calibration of building energy models. (2014). https://doi.org/10.13140/2.1.1674.9127

    2014

  8. [16]

    Bul- letin of Mathematical Biology79(4), 939–974 (2017) https://doi.org/10.1007/ s11538-017-0258-5

    Collis, J., Connor, A.J., Paczkowski, M., Kannan, P., Pitt-Francis, J., Byrne, 35 H.M., Hubbard, M.E.: Bayesian Calibration, Validation and Uncertainty Quan- tification for Predictive Modelling of Tumour Growth: A Tutorial. Bul- letin of Mathematical Biology79(4), 939–974 (201...

    2017

  9. [17]

    Archives of Computational Methods in Engineering30(5), 2859–2888 (2023) https://doi.org/10.1007/s11831-023-09888-y

    Pablos, J.L., Sabirov, I., Romero, I.: An Experimental, Computational, and Statistical Strategy for the Bayesian Calibration of Complex Material Models. Archives of Computational Methods in Engineering30(5), 2859–2888 (2023) https://doi.org/10.1007/s11831-023-09888-y . Accesse...

    2023 doi

  10. [18]

    Materials13(19) (2020) https://doi

    Pablos, J.L., Menga, E., Romero, I.: A Methodology for the Statistical Cali- bration of Complex Constitutive Material Models: Application to Temperature- Dependent Elasto-Visco-Plastic Materials. Materials13(19) (2020) https://doi. org/10.3390/ma13194402 . Accessed 2023-11-03

    2020 doi

  11. [19]

    Geosci- entific Model Development18(20), 7501–7527 (2025) https://doi.org/10.5194/ gmd-18-7501-2025

    Beylat, S., Raoult, N., Bacour, C., Douglas, N., Quaife, T., Bastrikov, V., Rayner, P.J., Peylin, P.: Towards the assimilation of atmospheric CO 2 concentration data in a land surface model using adjoint-free variational methods. Geosci- entific Model Development18(20), 7501–7...

    2025

  12. [20]

    Journal of Machine Learning Research24, 1–35 (2023)

    Spitieris, M., Steinsland, I.: Bayesian calibration of imperfect computer models using physics-informed priors. Journal of Machine Learning Research24, 1–35 (2023)

    2023

  13. [21]

    Preprint, University of Luxembourg (2024)

    Hennicker, J., Gansen, A., Sill, C., Verdonc, N., Baller, J., Hale, J.S.: Bayesian calibration of a model of polymer die swell using data from a laser–based mea- surement system. Preprint, University of Luxembourg (2024). https://orbilu.uni. lu/bitstream/10993/62576/1/hennicke...

    2024

  14. [22]

    Com- puters & Chemical Engineering179, 108430 (2023) https://doi.org/10.1016/j

    Eugene, E.A., Jones, K.D., Gao, X., Wang, J., Dowling, A.W.: Learning and optimization under epistemic uncertainty with Bayesian hybrid models. Com- puters & Chemical Engineering179, 108430 (2023) https://doi.org/10.1016/j. compchemeng.2023.108430

    2023

  15. [23]

    https://arxiv.org/abs/1206.5015

    Ling, Y., Mullins, J., Mahadevan, S.: Calibration of multi-physics computational models using Bayesian networks (2012). https://arxiv.org/abs/1206.5015

    2012 arXiv

  16. [24]

    Bayesian Analysis, 1–30 (2022) https://doi.org/10.1214/21-BA1293

    Marmin, S., Filippone, M.: Deep Gaussian Processes for Calibration of Computer Models. Bayesian Analysis, 1–30 (2022) https://doi.org/10.1214/21-BA1293

    2022 doi

  17. [25]

    In: Biegler, L.T., Biros, G., Ghattas, O., Heinkenschloss, M., Keyes, D., Mallick, B.K., Tenorio, L., Bloemen Waanders, B., Willcox, K

    Wilkinson, R.D.: Bayesian calibration of expensive multivariate computer exper- iments. In: Biegler, L.T., Biros, G., Ghattas, O., Heinkenschloss, M., Keyes, D., Mallick, B.K., Tenorio, L., Bloemen Waanders, B., Willcox, K. (eds.) Large- Scale Inverse Problems and Quantificati...

    2010 doi

  18. [26]

    arXiv preprint arXiv:1808.01932 (2018)

    Carmassi, M., Barbillon, P., Chiodetti, M., Keller, M., Parent, E.: Calico: a R package for Bayesian calibration. arXiv preprint arXiv:1808.01932 (2018)

    2018 arXiv

  19. [27]

    Software (2018)

    Carmassi, M.: CaliCo: Bayesian Calibration for Computational Codes. Software (2018). https://github.com/Demiperimetre/CaliCo

    2018

  20. [28]

    Virgo, N., Kleywegt, A., Merwe, D.: Calibrator: Bayesian Calibration of Com- puter Models. (2019). R package version 1.0. https://cran.r-project.org/package= calibrator

    2019

  21. [29]

    PeerJ Computer Science2, 55 (2016) https://doi.org/10.7717/ peerj-cs.55

    Salvatier, J., Wiecki, T.V., Fonnesbeck, C.: Probabilistic programming in Python using PyMC3. PeerJ Computer Science2, 55 (2016) https://doi.org/10.7717/ peerj-cs.55

    2016

  22. [30]

    https://github.com/jamesbriant/KOH-GPJax (2023)

    Briant, J.: KOH-GPJax: Kennedy & O’Hagan (2001) Gaussian Process Calibra- tion. https://github.com/jamesbriant/KOH-GPJax (2023)

    2001

  23. [31]

    https://github

    Idaho National Laboratory: BayCal: Bayesian calibration toolkit. https://github. com/idaholab/BayCal (2022)

    2022

  24. [32]

    Preprint, referenced as of 2025 (2015)

    Deardon, R., Goldstein, M., Forest, C.E., et al.: QUEENS: An Open-Source Python Framework for Solver Uncertainty Quantification. Preprint, referenced as of 2025 (2015). https://arxiv.org/abs/1507.05089

    2025 arXiv

  25. [33]

    arXiv preprint arXiv:2508.16316 (2025)

    Biehler, J., Nitzler, J., Brandst¨ ater, S., Dinkel, M., Gravemeier, V., H¨ ausel, L.J., Robalo Rei, G., Willmann, H., Wirthl, B., Wall, W.A.: Queens: An open-source Python framework for solver-independent analyses of large-scale computational models. arXiv preprint arXiv:2508...

    2025 arXiv

  26. [34]

    GitLab (2026)

    Schenk, C., Romero, I.: ACBICI: A Configurable Bayesian Calibration and Inference Package. GitLab (2026). https://gitlab.com/schenkch/ACBICI

    2026

  27. [35]

    ReadtheDocs (2026)

    Schenk, C., Romero, I.: Documentation for ACBICI: A Configurable Bayesian Calibration and Inference Package. ReadtheDocs (2026). https://acbici. readthedocs.io

    2026

  28. [36]

    Bayesian Analysis1(4), 765–792 (2006)

    Williams, B., Higdon, D., Gattiker, J., Moore, L., McKay, M., Keller-McNulty, S.: Combining Experimental Data and Computer Simulations, With an Application to Flyer Plate Experiments. Bayesian Analysis1(4), 765–792 (2006)

    2006

  29. [37]

    Tree Physiology25(7), 915– 927 (2005) https://doi.org/10.1093/treephys/25.7.915

    Van Oijen, M., Rougier, J., Smith, R.: Bayesian calibration of process-based forest models: bridging the gap between models and data. Tree Physiology25(7), 915– 927 (2005) https://doi.org/10.1093/treephys/25.7.915

    2005 doi

  30. [38]

    McClarren, R.G.: Predictive Models Informed by Simulation, Measurement, and Surrogates, pp. 275–304. Springer, Cham (2018). Chap. 11

    2018

  31. [39]

    SIAM/ASA Journal on Uncertainty Quantification6(1), 151–179 (2018) https://doi.org/10.1137/ 15m1033162

    Damblin, G., Barbillon, P., Keller, M., Pasanisi, A., Parent, E.: Adaptive 37 numerical designs for the calibration of computer codes. SIAM/ASA Journal on Uncertainty Quantification6(1), 151–179 (2018) https://doi.org/10.1137/ 15m1033162

    2018

  32. [40]

    Nature Reviews Methods Primers1(1) (2021) https://doi.org/10

    Schoot, R., Depaoli, S., King, R., Kramer, B., M¨ artens, K., Tadesse, M.G., Van- nucci, M., Gelman, A., Veen, D., Willemsen, J., Yau, C.: Bayesian statistics and modelling. Nature Reviews Methods Primers1(1) (2021) https://doi.org/10. 1038/s43586-020-00001-2

    2021

  33. [41]

    Mathemat- ical Modeling and Computational Experiment1(4), 407–414 (1993)

    Sobol, I.M.: Sensitivity estimates for nonlinear mathematical models. Mathemat- ical Modeling and Computational Experiment1(4), 407–414 (1993). Translated from Russian

    1993

  34. [42]

    Society for Industrial and Applied Mathemat- ics, Philadelphia, PA (2024)

    Smith, R.C.: Uncertainty Quantification: Theory, Implementation, and Applications, 2nd edn. Society for Industrial and Applied Mathemat- ics, Philadelphia, PA (2024). https://doi.org/10.1137/1.9781611977844 . https://epubs.siam.org/doi/abs/10.1137/1.9781611977844

    2024 doi

  35. [43]

    Polymer Reaction Engineering11(3), 563–588 (2003)

    Yao, K.Z., Shaw, B.M., Kou, B., McAuley, K.B., Bacon, D.: Modeling ethylene/butene copolymerization with multi-site catalysts: parameter estima- bility and experimental design. Polymer Reaction Engineering11(3), 563–588 (2003)

    2003

  36. [44]

    Wu, S., McLean, K.A.P., Harris, T.J., McAuley, K.B.: Selection of opti- mal parameter set using estimability analysis and mse-based model- selection criterion. International Journal of Advanced Mechatronic Sys- tems3(3), 188–197 (2011) https://doi.org/10.1504/IJAMECHS.2011.042...

    2011 arXiv

  37. [45]

    Com- puters & Chemical Engineering134, 106716 (2020) https://doi.org/10.1016/j

    Schenk, C., Short, M., Rodriguez, J.S., Thierry, D., Biegler, L.T., Garc ´ ıa-Mu˜ noz, S., Chen, W.: Introducing kipet: A novel open-source software package for kinetic parameter estimation from experimental datasets including spectra. Com- puters & Chemical Engineering134, 10...

    2020

  38. [46]

    Fields Institute Monographs

    Madras, N.: Lectures on Monte Carlo Methods. Fields Institute Monographs. American Mathematical Society, Providence, Rhode Island (2002)

    2002

  39. [47]

    Springer Texts in Statistics

    Robert, C.P., Casella, G.: Monte Carlo Statistical Methods. Springer Texts in Statistics. Springer, New York (2004). https://doi.org/10.1007/ 978-1-4757-4145-2 .http://dx.doi.org/10.1007/978-1-4757-4145-2

    2004 doi

  40. [48]

    Springer, Singapore (2020)

    Barbu, A., Zhu, S.-C.: Monte Carlo Methods. Springer, Singapore (2020)

    2020

  41. [49]

    Mecca- nica60, 1411–1436 (2025) https://doi.org/10.1007/s11012-025-01997-1 38

    Romero, I., Ortiz, M.: An energy-stepping Markov Monte Carlo method. Mecca- nica60, 1411–1436 (2025) https://doi.org/10.1007/s11012-025-01997-1 38

    2025 doi

  42. [50]

    Springer series in statistics

    Hastie, T., Tibshirani, R., Friedman, J.: The Elements of Statistical Learning: Data Mining, Inference, and Prediction. Springer series in statistics. Springer, New York, NY (2009)

    2009

  43. [51]

    Machine Learning45(1), 5–32 (2001) https://doi

    Breiman, L.: Random forests. Machine Learning45(1), 5–32 (2001) https://doi. org/10.1023/a:1010933404324

    2001 doi

  44. [52]

    MIT Press, Cambridge, Massachusetts (2006)

    Rasmussen, C.E., Williams, C.K.I.: Gaussian Processes for Machine Learning. MIT Press, Cambridge, Massachusetts (2006)

    2006

  45. [53]

    Technometrics21(2), 55–61 (1979) https://doi.org/10.2307/1268522

    McKay, M., Beckman, R., Conover, W.: A comparison of three methods for select- ing values of input variables in the analysis of output from a computer code. Technometrics21(2), 55–61 (1979) https://doi.org/10.2307/1268522

    1979 doi

  46. [54]

    Journal of Quality Technology13(3), 174–183 (1981) https: //doi.org/10.1080/00224065.1981.11978748

    Iman, R., Helton, J., Campbell, J.: An approach to sensitivity analysis of com- puter models: Part i—introduction, input variable selection and preliminary variable assessment. Journal of Quality Technology13(3), 174–183 (1981) https: //doi.org/10.1080/00224065.1981.11978748

    1981

  47. [55]

    In: 2023 IEEE International Conference on Robotics and Automation (ICRA), pp

    Esposito, J.M.: Concentration of measure phenomenon and its implications for sample-based planning algorithms in very-high dimensional configuration spaces. In: 2023 IEEE International Conference on Robotics and Automation (ICRA), pp. 7865–7871 (2023). https://doi.org/10.1109/...

    2023

  48. [56]

    Computational Materials Science252, 113780 (2025)

    Schenk, C., Haranczyk, M.: A novel constrained sampling method for efficient exploration in materials and chemical mixture design. Computational Materials Science252, 113780 (2025)

    2025

  49. [57]

    Chemometrics and Intelligent Laboratory Systems230, 104683 (2022) https://doi.org/10.1016/j.chemolab

    Bashiri, S., Yasari, E., Tayyebi, S.: Comparison of different sampling and surro- gate modelling approaches for a multi-objective optimization problem of direct dimethyl ether synthesis in the fixed-bed reactor. Chemometrics and Intelligent Laboratory Systems230, 104683 (2022)...

    2022

  50. [58]

    Machine Learning with Applications9, 100373 (2022) https://doi.org/10.1016/j.mlwa.2022.100373

    Kamath, C.: Intelligent sampling for surrogate modeling, hyperparameter opti- mization, and data analysis. Machine Learning with Applications9, 100373 (2022) https://doi.org/10.1016/j.mlwa.2022.100373 . Accessed 2026-01-08

    2022

  51. [59]

    In: Adams, R.P., Gogate, V

    Wu, J., Toscano-Palmerin, S., Frazier, P.I., Wilson, A.G.: Practical multi-fidelity Bayesian optimization for hyperparameter tuning. In: Adams, R.P., Gogate, V. (eds.) Proceedings of The 35th Uncertainty in Artificial Intelligence Conference. Proceedings of Machine Learning Re...

    2020

  52. [60]

    In: Ruiz, F., Dy, J., Meent, J.-W

    Mikkola, P., Martinelli, J., Filstroff, L., Kaski, S.: Multi-fidelity Bayesian opti- mization with unreliable information sources. In: Ruiz, F., Dy, J., Meent, J.-W. (eds.) Proceedings of The 26th International Conference on Artificial Intelligence 39 and Statistics. Proceedin...

    2023

  53. [61]

    Dig- ital Discovery2(6), 1937–1956 (2023) https://doi.org/10.1039/D3DD00117B

    Gantzler, N., Deshwal, A., Doppa, J.R., Simon, C.M.: Multi-fidelity Bayesian optimization of covalent organic frameworks for xenon/krypton separations. Dig- ital Discovery2(6), 1937–1956 (2023) https://doi.org/10.1039/D3DD00117B . Accessed 2024-04-23

    1937 doi

  54. [62]

    Advanced Intelligent Discovery, 202500054 https://doi.org/10.1002/aidi

    Ozdemir, B., Hern´ andez-del-Valle, M., Schenk, C., Wang, D.-Y., Haranczyk, M.: Bayesian optimization guiding the experimental mapping of the Pareto front of mechanical and flame-retardant properties in polyamide nanocompos- ites. Advanced Intelligent Discovery, 202500054 http...

    doi

  55. [63]

    Publications of the Astronomical Society of the Pacific125(925), 306– 312 (2013) https://doi.org/10.1086/670067

    Foreman-Mackey, D., Hogg, D.W., Lang, D., Goodman, J.: emcee: The MCMC Hammer. Publications of the Astronomical Society of the Pacific125(925), 306– 312 (2013) https://doi.org/10.1086/670067 . arXiv:1202.3665 [astro-ph]. Accessed 2025-12-09

    2013 arXiv

  56. [64]

    In: Larochelle, H., Ranzato, M., Hadsell, R., Balcan, M.F., Lin, H

    Acerbi, L.: Variational Bayesian Monte Carlo with noisy likelihoods. In: Larochelle, H., Ranzato, M., Hadsell, R., Balcan, M.F., Lin, H. (eds.) Advances in Neural Information Processing Systems, vol. 33, pp. 8211–8222 (2020)

    2020

  57. [65]

    Journal of Open Source Software8(86), 5428 (2023) https://doi.org/10.21105/joss.05428

    Huggins, B., Li, C., Tobaben, M., Aarnos, M.J., Acerbi, L.: Pyvbmc: Efficient Bayesian inference in Python. Journal of Open Source Software8(86), 5428 (2023) https://doi.org/10.21105/joss.05428

    2023 doi

  58. [66]

    In: Bengio, S., Wallach, H., Larochelle, H., Grauman, K., Cesa-Bianchi, N., Garnett, R

    Acerbi, L.: Variational Bayesian Monte Carlo. In: Bengio, S., Wallach, H., Larochelle, H., Grauman, K., Cesa-Bianchi, N., Garnett, R. (eds.) Advances in Neural Information Processing Systems, vol. 31 (2018)

    2018

  59. [67]

    International Journal for Numerical Methods in Engineering121(5), 904–924 (2019) https://doi.org/10

    Menga, E., S´ anchez, M.J., Romero, I.: Anisotropic meta-models for computa- tionally expensive simulations in nonlinear mechanics. International Journal for Numerical Methods in Engineering121(5), 904–924 (2019) https://doi.org/10. 1002/nme.6250

    2019

  60. [68]

    Scientific Reports 10(1), 17663 (2020) https://doi.org/10.1038/s41598-020-74394-1

    Noack, M.M., Doerk, G.S., Li, R., Streit, J.K., Vaia, R.A., Yager, K.G., Fukuto, M.: Autonomous materials discovery driven by Gaussian process regression with inhomogeneous measurement noise and anisotropic kernels. Scientific Reports 10(1), 17663 (2020) https://doi.org/10.103...

    2020 doi

  61. [69]

    Dig- ital Discovery2(6), 1969–1979 (2023) https://doi.org/10.1039/D3DD00141E

    Hern´ andez-del-Valle, M., Schenk, C., Echevarr ´ ıa-Pastrana, L., Ozdemir, B., Dios-L´ azaro, E., Ilarraza-Zuazo, J., Wang, D.-Y., Haranczyk, M.: Robotically automated 3D printing and testing of thermoplastic material specimens. Dig- ital Discovery2(6), 1969–1979 (2023) https...

    1969 doi

  62. [70]

    Communi- cations in Applied Mathematics and Computational Science5(1), 65–80 (2010) https://doi.org/10.2140/camcos.2010.5.65

    Goodman, J., Weare, J.: Ensemble samplers with affine invariance. Communi- cations in Applied Mathematics and Computational Science5(1), 65–80 (2010) https://doi.org/10.2140/camcos.2010.5.65 . Accessed 2026-01-08

    2010 doi

  63. [71]

    Signal Processing131, 386–401 (2017) https://doi.org/10.1016/j.sigpro.2016.08.025

    Martino, L., Elvira, V., Louzada, F.: Effective Sample Size for Importance Sam- pling based on discrepancy measures. Signal Processing131, 386–401 (2017) https://doi.org/10.1016/j.sigpro.2016.08.025 . arXiv:1602.03572 [stat]. Accessed 2026-01-08

    2017 arXiv

  64. [72]

    Statistical Science7(4), 457–472 (1992) https://doi.org/10.1214/ss/ 1177011136

    Gelman, A., Rubin, D.B.: Inference from Iterative Simulation Using Multiple Sequences. Statistical Science7(4), 457–472 (1992) https://doi.org/10.1214/ss/ 1177011136 . Publisher: Institute of Mathematical Statistics. Accessed 2026-01-08

    1992 doi

  65. [73]

    Statistical Science36(4), 518–529 (2021) https://doi.org/10.1214/20-STS812

    Vats, D., Knudson, C.: Revisiting the Gelman–Rubin Diagnostic. Statistical Science36(4), 518–529 (2021) https://doi.org/10.1214/20-STS812 . Publisher: Institute of Mathematical Statistics. Accessed 2026-01-08

    2021 doi

  66. [74]

    Bayesian Analysis16(2) (2021) https://doi.org/10

    Vehtari, A., Gelman, A., Simpson, D., Carpenter, B., B¨ urkner, P.-C.: Rank- normalization, folding, and localization: An improved ˆRfor assessing convergence of MCMC (with discussion). Bayesian Analysis16(2) (2021) https://doi.org/10. 1214/20-ba1221

    2021

  67. [75]

    The American Eco- nomic Review18(1), 139–165 (1928)

    Cobb, C.W., Douglas, P.H.: A Theory of Production. The American Eco- nomic Review18(1), 139–165 (1928). Publisher: American Economic Association. Accessed 2026-01-08

    1928

  68. [76]

    data/experiments.dat

    Barro, R.J., Sala-i-Martin, X.: Economic Growth, 2nd ed edn. MIT Press, Cambridge, Mass (2004) 41 Supplementary Information Gravity example - full scripts and complete MCMC and VBMC diagnostics This example illustrates the calibration of a simple physical model that has been p...

    2004

Pith tools

Reviewed August 3, 2026 · model on record in the stance chip above.