Abstract:It is often desirable to build a statistical emulator of a complex computer simulator in order to perform analysis which would otherwise be computationally infeasible. We propose methodology to model multivariate output from a computer simulator taking into account output structure in the responses. The utility of this approach is demonstrated by applying it to a chemical and biological hazard prediction model. Predicting the hazard area which results from an accidental or deliberate chemical or biological release is imperative in civil and military planning and also in emergency response. The hazard area resulting from such a release is highly structured in space, and we therefore propose the use of a thin-plate spline to capture the spatial structure and fit a Gaussian process emulator to the coefficients of the resultant basis functions. We compare and contrast four different techniques for emulating multivariate output: dimension reduction using (i) a fully Bayesian approach with a principal component basis, (ii) a fully Bayesian approach with a thin-plate spline basis, assuming that the basis coefficients are independent, and (iii) a plug-in Bayesian approach with a thin-plate spline basis and a separable covariance structure; and (iv) a functional data modeling approach using a tensor-product (separable) Gaussian process. We develop methodology for the two thin-plate spline emulators and demonstrate that these emulators significantly outperform the principal component emulator. Further, the separable thin-plate spline emulator, which accounts for the dependence between basis coefficients, provides substantially more realistic quantification of uncertainty, and is also computationally more tractable, allowing fast emulation. For high resolution output data, it also offers substantial predictive and computational advantages over the tensor-product Gaussian process emulator.

A Statistical Method for Emulation of Computer Models With Invariance-Preserving Properties, With Application to Structural Energy Prediction

Neural Network Emulation of Spatio-temporal Data Using Linear and Nonlinear Dimensionality Reduction

Machine learning strategies for systems with invariance properties

A Fast and Calibrated Computer Model Emulator: An Empirical Bayes Approach

Embed and Emulate: Contrastive representations for simulation-based inference

Model-free Data-Driven inference in computational mechanics

Efficient Emulation of Computer Models Utilising Multiple Known Boundaries of Differing Dimensions

Similarity Equivariant Graph Neural Networks for Homogenization of Metamaterials

Statistical finite elements for misspecified models

Thermodynamic Fidelity of Generative Models for Ising System

Inferring Displacement Fields from Sparse Measurements Using the Statistical Finite Element Method

Reliable emulation of complex functionals by active learning with error control

Emulation of Multivariate Simulators Using Thin-Plate Splines with Application to Atmospheric Dispersion

Efficient Surrogate Models for Materials Science Simulations: Machine Learning-based Prediction of Microstructure Properties

Low-dimensional Data-based Surrogate Model of a Continuum-mechanical Musculoskeletal System Based on Non-intrusive Model Order Reduction

InVAErt networks: A data-driven framework for model synthesis and identifiability analysis

Functional order-reduced Gaussian Processes based machine-learning emulators for probabilistic constitutive modelling

The Stochastic Replica Approach to Machine Learning: Stability and Parameter Optimization

Emulating Expert Insight: A Robust Strategy for Optimal Experimental Design

Surrogate recycling for structures with spatially uncertain stiffness

Emulation methods and adaptive sampling increase the efficiency of sensitivity analysis for computationally expensive models