Adaptive design for Gaussian process regression under censoring

DOI10.1214/21-AOAS1512zbMATH Open1498.62153arXiv1910.05452OpenAlexW2979672915WikidataQ114060490 ScholiaQ114060490MaRDI QIDQ2154177FDOQ2154177

Authors: Jialei Chen, Simon Mak, V. Roshan Joseph, Chuck Zhang 

Publication date: 14 July 2022

Published in: The Annals of Applied Statistics (Search for Journal in Brave)

Abstract: A key objective in engineering problems is to predict an unknown experimental surface over an input domain. In complex physical experiments, this may be hampered by response censoring, which results in a significant loss of information. For such problems, experimental design is paramount for maximizing predictive power using a small number of expensive experimental runs. To tackle this, we propose a novel adaptive design method, called the integrated censored mean-squared error (ICMSE) method. The ICMSE method first estimates the posterior probability of a new observation being censored, then adaptively chooses design points that minimize predictive uncertainty under censoring. Adopting a Gaussian process regression model with product correlation function, the proposed ICMSE criterion is easy to evaluate, which allows for efficient design optimization. We demonstrate the effectiveness of the ICMSE design in two real-world applications on surgical planning and wafer manufacturing.

Full work available at URL: https://arxiv.org/abs/1910.05452

Recommendations

• A discussion on adaptive designs for computer experiments
• Optimal experimental designs when an independent variable is potentially censored
• Robust prediction and extrapolation designs for censored data
• Construction and efficient implementation of adaptive objective-based designs of experiments
• scientific article; zbMATH DE number 5323623

zbMATH Keywords

experimental designkrigingadaptive samplingmultifidelity modelingcensored experiments

Mathematics Subject Classification ID

Computational methods for problems pertaining to statistics (62-08) Bayesian inference (62F15) Applications of statistics in engineering and industry; control charts (62P30) Gaussian processes (60G15) Design of statistical experiments (62K99)

Cites Work

• Strictly Proper Scoring Rules, Prediction, and Estimation
• Design and analysis of computer experiments. With comments and a rejoinder by the authors
• Bayesian calibration of computer models. (With discussion)
• Stochastic kriging for simulation metamodeling
• A Simplex Method for Function Minimization
• Efficient emulators of computer experiments using compactly supported correlation functions, with an application to cosmology
• A supermartingale approach to Gaussian process based sequential design of experiments
• Maximum projection designs for computer experiments
• On the distribution of points in a cube and the approximate evaluation of integrals
• The design and analysis of computer experiments
• An efficient surrogate model for emulation and physics extraction of large eddy simulations
• Exploratory designs for computational experiments
• Gaussian process modeling with inequality constraints
• Finite-Dimensional Gaussian Approximation with Linear Inequality Constraints

Cited In (3)

• A review on computer model calibration
• Targeted adaptive design
• Adversarial Gaussian Process Regression in Sensor Networks

Uses Software

• COMSOL

This page was built for publication: Adaptive design for Gaussian process regression under censoring

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2154177)