Files in this item

FilesDescriptionFormat

application/pdf

application/pdfECE499-Sp2019-kim.pdf (2MB)Restricted to U of Illinois
(no description provided)PDF

Description

Title:Exploration of acquisition functions in Bayesian optimization and their applications in electrical engineering
Author(s):Kim, Kevin
Contributor(s):Schutt-Ainé, José E.
Subject(s):Bayesian Optimization
Acquisition Functions
Gaussian Process
Global Optimization
Abstract:Bayesian Optimization (BO) is commonly used for globally optimizing black-box functions. In short, Bayesian Optimization uses a probabilistic model of the black-box function and a surrogate function known as the acquisition function is built upon this model. This function guides the maximization process by describing the expected utility of performing an evaluation of the black-box function at a given point. Different measures of “expected utility” creates different closed-forms of the acquisition function which ultimately causes different convergence rates to the global optimum of the objective function. This work compares the convergence rate using different acquisition functions proposed in literature on synthetic functions and on electrical engineering examples.
Issue Date:2019-05
Genre:Other
Type:Text
Language:English
URI:http://hdl.handle.net/2142/104021
Date Available in IDEALS:2019-06-14


This item appears in the following Collection(s)

Item Statistics