arrow
返回

A Hierarchical Expected Improvement Method for Bayesian Optimization

delete2023-06-23
delete12
delete
OA
AI
Z
Zhehui Chen
S
Simon Mak
C
Changbao Wu *
DOI:10.1080/01621459.2023.2210803delete
delete原文链接
delete原文求助
delete分享
delete收藏
摘要

摘要

En 中文
The Expected Improvement (EI) method, proposed by Jones, Schonlau, andWelch, is a widely used Bayesian optimization method, which makes use of a fitted Gaussian process model for efficient black-box optimization. However, one key drawback of EI is that it is overly greedy in exploiting the fitted Gaussian process model for optimization, which results in suboptimal solutions even with large sample sizes. To address this, we propose a new hierarchical EI (HEI) framework, which makes use of a hierarchical Gaussian process model. HEI preserves a closed-form acquisition function, and corrects the over-greediness of EI by encouraging exploration of the optimization space. We then introduce hyperparameter estimation methods which allow HEI to mimic a fully Bayesian optimization procedure, while avoiding expensive Markov-chain Monte Carlo sampling steps. We prove the global convergence of HEI over a broad function space, and establish near-minimax convergence rates under certain prior specifications. Numerical experiments show the improvement of HEI over existing Bayesian optimization methods, for synthetic functions and a semiconductor manufacturing optimization problem. for this article are available online.
Keyword:
Black-box optimization
Experimental design
Gaussian process
Global optimization
Hierarchical modeling
Uncertainty quantification

期刊

J
Journal of the American Statistical Association
IF:
3
论文数:
5.2K
被引数:
4.8W

机构

G
Georgia Institute of Technology
学者数:
1.8W
论文数: 1.4W
被引数: 5.9W
U
university system of georgia
学者数:
7.3W
论文数: 6.5W
被引数: 101
引用论文

引用论文

Statistical modeling and analysis for robust synthesis of nanostructures
err2012-01-01
err39
PREAI
errDasgupta, Tirthankar; Ma, Christopher; Joseph, V. Roshan; Wang, Z. L.; Wu, C. F. Jeff
err分享
err收藏
Utilization of phage display to identify antigenic regions in the PCV2 capsid protein for the evaluation of serological responses in mice and pigs利用噬菌体展示技术识别PCV2衣壳蛋白中的抗原决定簇区域,以评估小鼠和猪的血清学反应
err2018-03-27
err0
errOAAI
errMarcus Rebouças Santos; Viviane Sisdelli Assao; Fabiana de Almeida Araújo Santos; Rafael Locatelli Salgado; Ana Paula Carneiro; Juliana Lopes Rangel Fietto; Gustavo Costa Bressan; Márcia Rogéria de Almeida; Zelia Inês Portela Lobato; Carlos Ueira-Veira; Luíz Ricardo Goulart; Abelardo Silva-Júnior
err分享
err收藏
err分享
err收藏
Examining the sources of variability in cell culture media used for biopharmaceutical production
err2017-09-22
err0
PREAI
errNicola McGillicuddy; Patrick Floris; Simone Albrecht; Jonathan Bones
err分享
err收藏
Interventions to Reduce Rehospitalizations after Chronic Obstructive Pulmonary Disease Exacerbations. A Systematic Review
err2014-03-01
err0
errOAAI
errValentin Prieto-Centurion; Michael A. Markos; Norma I. Ramey; Hélène A. Gussin; Sharmilee M. Nyenhuis; Min J. Joo; Bharati Prasad; Nina Bracken; Robert DiDomenico; Patrick O. Godwin; Howard A. Jaffe; Ravi Kalhan; Alan S. Pickard; Barry R. Pittendrigh; Bruce Schatz; Jamie L. Sullivan; Byron M. Thomashow; Mark V. Williams; Jerry A. Krishnan
err分享
err收藏
BAYESIAN LINEAR REGRESSION WITH SPARSE PRIORS
err2015-10-01
err284
errOAAI
errCastillo, Ismael; Schmidt-Hieber, Johannes; Van der Vaart, Aad
err分享
err收藏
学者 查看更多内容