Locally induced Gaussian processes for large-scale simulation experiments
From MaRDI portal
Publication:2058747
DOI10.1007/S11222-021-10007-9zbMATH Open1475.62023arXiv2008.12857OpenAlexW3154293753MaRDI QIDQ2058747FDOQ2058747
Ryan B. Christianson, D. Austin Cole, Robert B. Gramacy
Publication date: 9 December 2021
Published in: Statistics and Computing (Search for Journal in Brave)
Abstract: Gaussian processes (GPs) serve as flexible surrogates for complex surfaces, but buckle under the cubic cost of matrix decompositions with big training data sizes. Geospatial and machine learning communities suggest pseudo-inputs, or inducing points, as one strategy to obtain an approximation easing that computational burden. However, we show how placement of inducing points and their multitude can be thwarted by pathologies, especially in large-scale dynamic response surface modeling tasks. As remedy, we suggest porting the inducing point idea, which is usually applied globally, over to a more local context where selection is both easier and faster. In this way, our proposed methodology hybridizes global inducing point and data subset-based local GP approximation. A cascade of strategies for planning the selection of local inducing points is provided, and comparisons are drawn to related methodology with emphasis on computer surrogate modeling applications. We show that local inducing points extend their global and data-subset component parts on the accuracy--computational efficiency frontier. Illustrative examples are provided on benchmark data and a large-scale real-simulation satellite drag interpolation problem.
Full work available at URL: https://arxiv.org/abs/2008.12857
Cites Work
- Design and analysis of computer experiments. With comments and a rejoinder by the authors
- Bayesian Treed Gaussian Process Models With an Application to Computer Modeling
- Stochastic Kriging for Simulation Metamodeling
- Regularization Algorithms for Learning That Are Equivalent to Multilayer Networks
- Title not available (Why is that?)
- Gaussian Predictive Process Models for Large Spatial Data Sets
- Approximating Likelihoods for Large Spatial Data Sets
- A Limited Memory Algorithm for Bound Constrained Optimization
- Title not available (Why is that?)
- A general framework for Vecchia approximations of Gaussian processes
- Sparse on-line Gaussian processes
- Title not available (Why is that?)
- A Comparison of Three Methods for Selecting Values of Input Variables in the Analysis of Output from a Computer Code
- Title not available (Why is that?)
- Title not available (Why is that?)
- Title not available (Why is that?)
- Knot selection in sparse Gaussian processes with a variational objective function
- Emulating Satellite Drag from Large Simulation Experiments
- Analyzing Nonstationary Spatial Data Using Piecewise Gaussian Processes
- Efficient emulators of computer experiments using compactly supported correlation functions, with an application to cosmology
- Hilbert space methods for reduced-rank Gaussian process regression
- Quantifying Uncertainties on Excursion Sets Under a Gaussian Random Field Prior
- A supermartingale approach to Gaussian process based sequential design of experiments
- Massively Parallel Approximate Gaussian Process Regression
- Exploiting Variance Reduction Potential in Local Gaussian Process Search
- The design and analysis of computer experiments
- Exploratory designs for computational experiments
- Variational inference for sparse spectrum Gaussian process regression
- Title not available (Why is that?)
- Title not available (Why is that?)
- Recursive estimation for sparse Gaussian process regression
- Title not available (Why is that?)
- Parameter estimation in high dimensional Gaussian distributions
- Spectral Approximation of the IMSE Criterion for Optimal Designs in Kernel-Based Interpolation Models
- Computer experiment designs for accurate prediction
- Fast Estimation of $tr(f(A))$ via Stochastic Lanczos Quadrature
- Adaptive Gaussian Process Approximation for Bayesian Inference with Expensive Likelihood Functions
- Mercer Kernels and Integrated Variance Experimental Design: Connections Between Gaussian Process Regression and Polynomial Approximation
Cited In (9)
- Active Learning for Deep Gaussian Process Surrogates
- Real-Time Local GP Model Learning
- Batch-sequential design and heteroskedastic surrogate modeling for delta smelt conservation
- Large-scale local surrogate modeling of stochastic simulation experiments
- A Global-Local Approximation Framework for Large-Scale Gaussian Process Modeling
- Sensitivity Prewarping for Local Surrogate Modeling
- Augmenting a Simulation Campaign for Hybrid Computer Model and Field Data Experiments
- Vecchia-approximated Deep Gaussian Processes for Computer Experiments
- Exploiting Variance Reduction Potential in Local Gaussian Process Search
This page was built for publication: Locally induced Gaussian processes for large-scale simulation experiments
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2058747)
