Sequential Bayesian Experimental Design for Calibration of Expensive Simulation Models

Sürer, Özge; Plumlee, Matthew; Wild, Stefan M.

doi:10.1080/00401706.2023.2246157

Cited by 5 publications

(3 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Williams et al (2011) explored entropy and distance-based criteria in a batch sequential design setting for the physical experiment by improving the global prediction of discrepancies inferred from computer model calibration. Sürer et al (2023) focuses on sampling the simulation input θ s i by proposing a sequential framework with a criterion for parameter selection that targets learning the posterior density of the parameters. Koermer et al (2023) focuses on selecting the simulation input x s i , θ s i À Á for building the GP surrogate model in the KOH setting using the IMSPE criterion, for which they derive a closed-form expression facilitating the optimization.…”

Section: Sequential Design/active Learningmentioning

confidence: 99%

“…Sürer et al (2023) focuses on sampling the simulation input

{\boldsymbol{\theta}}_i^s

by proposing a sequential framework with a criterion for parameter selection that targets learning the posterior density of the parameters. Koermer et al (2023) focuses on selecting the simulation input

\left({\mathbf{x}}_i^s,{\boldsymbol{\theta}}_i^s\right)

for building the GP surrogate model in the KOH setting using the IMSPE criterion, for which they derive a closed‐form expression facilitating the optimization.…”

Section: Experimental Designsmentioning

confidence: 99%

See 1 more Smart Citation

A review on computer model calibration

Sung,

Tuo

2024

WIREs Computational Stats

View full text Add to dashboard Cite

Model calibration is crucial for optimizing the performance of complex computer models across various disciplines. In the era of Industry 4.0, symbolizing rapid technological advancement through the integration of advanced digital technologies into industrial processes, model calibration plays a key role in advancing digital twin technology, ensuring alignment between digital representations and real‐world systems. This comprehensive review focuses on the Kennedy and O'Hagan (KOH) framework (Kennedy and O'Hagan, Journal of the Royal Statistical Society: Series B 2001; 63(3):425–464). In particular, we explore recent advancements addressing the challenges of the unidentifiability issue while accommodating model inadequacy within the KOH framework. In addition, we explore recent advancements in adapting the KOH framework to complex scenarios, including those involving multivariate outputs and functional calibration parameters. We also delve into experimental design strategies tailored to the unique demands of model calibration. By offering a comprehensive analysis of the KOH approach and its diverse applications, this review serves as a valuable resource for researchers and practitioners aiming to enhance the accuracy and reliability of their computer models.This article is categorized under: Statistical Models > Semiparametric Models Statistical Models > Simulation Models Statistical Models > Bayesian Models

show abstract

Section: Sequential Design/active Learningmentioning

confidence: 99%

“…Sürer et al (2023) focuses on sampling the simulation input

{\boldsymbol{\theta}}_i^s

\left({\mathbf{x}}_i^s,{\boldsymbol{\theta}}_i^s\right)

for building the GP surrogate model in the KOH setting using the IMSPE criterion, for which they derive a closed‐form expression facilitating the optimization.…”

Section: Experimental Designsmentioning

confidence: 99%

A review on computer model calibration

Sung,

Tuo

2024

WIREs Computational Stats

View full text Add to dashboard Cite

show abstract

“…• Optimization of variational algorithms on quantum computers (Liu et al, 2022) • Parallelization of the ParMOO solver for multiobjective simulation optimization problems (Chang & Wild, 2023) • Design of particle accelerators (A. Ferran Pousa et al, 2022;A. Ferran Pousa et al, 2023;Neveu et al, 2023) • Sequential Bayesian experimental design (Sürer et al, 2023) and Bayesian calibration (Chan et al, 2023) A selection of community-provided libEnsemble functions and workflows that users can build off is maintained in libEnsemble Community (2023).…”

Section: Representative Libensemble Use Casesmentioning

confidence: 99%

libEnsemble: A complete Python toolkit for dynamic ensembles of calculations

Hudson,

Larson,

Navarro

et al. 2023

JOSS

Self Cite

View full text Add to dashboard Cite

Almost all science and engineering applications eventually stop scaling: their runtime no longer decreases as available computational resources increase. Therefore, many applications will struggle to efficiently use emerging extreme-scale high-performance, parallel, and distributed systems. libEnsemble is a complete Python toolkit and workflow system for intelligently driving ensembles of experiments or simulations at massive scales. It enables and encourages multidisciplinary design, decision, and inference studies portably running on laptops, clusters, and supercomputers. Statement of NeedWhile a growing number of packages are aimed at workflows, relatively few focus on running dynamic ensembles of calculations on clusters and supercomputers. Dynamic ensembles are workflows of computations that are defined and steered based on intermediate results. Examples include determining simulation parameters using numerical optimization methods, machine learning techniques, or statistical calibration tools. In each of these examples, the ensemble members are typically simulations that use different parameters or data. Additional examples of applications that have used libEnsemble are surveyed in the Representative libEnsemble Use Cases section below.The target audience for libEnsemble includes scientists, engineers, and other researchers who stand to benefit from such dynamic workflows.

show abstract