Bayesian calibration of computer models

By Marc C. Kennedy, Anthony OʼHagan. 2001

4,378 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Control and Systems Engineering, Statistics and Probability

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Control and Systems Engineering

In Statistics and Probability

Citations in the corpus

Counterfactual Normalization: Proactively Addressing Dataset Shift and Improving Reliability Using Causal Mechanisms (2018, methodology): “We apply the proposed method by ﬁtting an additive structural equation for Y using the Bayesian calibration form of Kennedy and O’Hagan (2001): where δ ( · ) is a Gaussian process (GP) prior (with RBF kernel) on the discrepancy function since our linear regression model is likely misspeciﬁed.”
Uncertainty quantification and management in additive manufacturing: current status, needs, and opportunities (2017, methodology): “Awidely used approach for dealing with this kind of problem is the employment of model calibration approach under the Kennedy and O’Hagan framework (KOH framework) [101].”
Model identification for hydrological forecasting under uncertainty (2005, methodology): “Another approach, introduced by Kennedy and O’Hagan (2001), uses a model inadequacy function in a Bayesian context to account for structural uncertainty by developing a more complex error model.”
Transforming (1996, neutral): “Accordingly, the participants were told to express their views on teaching requirements, arrangements, and challenges for cinematography education by leveraging many virtual environments inside the metaverse.”
A review of methods to match building energy simulation models to measured data (2014, methodology): “Kennedy and O ’ Hagan [73] present a generic approach for the Bayesian calibration of computer models.”
History matching for exploring and reducing climate model parameter space using observations and a large perturbed physics ensemble (2013, methodology): “Equation (2) specifies that the set { f(x), x } and g are independent for all x (Kennedy and O’Hagan 2001; Sexton et al. 2011).”
Structured Bayesian Gaussian process latent variable model: applications to data-driven dimensionality reduction and high-dimensional inversion (2018, methodology): “rested in determining some quantity of interest, but only has access to indirect measurements that are the outcome of some physical process. Examples of inverse problems include parameter calibration [1, 2], subsurface ow [3, 4], ocean dynamics [5], remote sensing [6], seismic inversion [7], and oth”
An overview of methods to evaluate uncertainty of deterministic models in decision support (2015, methodology): “The simplest way to conduct a UA is to apply Monte Carlo (MC)methods inwhich configurations of model inputs are drawn randomly from their distribution, and the resulting set of model outputs can be seen as a random sample of the distribution of the output of interest (Kennedy and O'Hagan, 2001).”
A stochastic collocation approach to Bayesian inference in inverse problems (2009, background): “Several efforts at accelerating Bayesian inference in inverse problems have appeared in recent literature; these have relied largely on reductions or surrogates for the forward model [3, 14, 17, 24], or instead have sought more efﬁcient sampling from the posterior [4,5,11].”
The end of model democracy? (2010, background): “there are processes that are not adequately resolved or parameterized, or missing entirely (Kennedy and O’Hagan 2001; McWilliams 2007; Smith 2002).”
The Challenge of Machine Learning in Space Weather: Nowcasting and Forecasting (2019, methodology): “The standard procedure, for high-dimensional problems (i.e., large number of parameters) is to resort to a Markov chain Monte Carlo (MCMC) approach (Gelman et al., 2013; Kennedy & O'Hagan, 2001).”
Quantifying and reducing model-form uncertainties in Reynolds-averaged Navier-Stokes simulations: A data-driven, physics-informed Bayesian approach (2015, methodology): “shared by all linear eddy-viscosity models that are commonly used in engineering practice, including the k{", k{!, and eddy-viscosity transport models. In their seminal work, Kennedy and O’Hagan [5] developed a Bayesian calibration approach that includes a model discrepancy term to account for ”