No Cover Image

Conference Paper/Proceeding/Abstract 724 views 208 downloads

On Bayesian Search for the Feasible Space Under Computationally Expensive Constraints

Alma Rahat Orcid Logo, Michael Wood

Machine Learning, Optimization, and Data Science, Volume: 12566, Pages: 529 - 540

Swansea University Author: Alma Rahat Orcid Logo

Check full text

DOI (Published version): 10.1007/978-3-030-64580-9_44

Abstract

We are often interested in identifying the feasible subset of a decision space under multiple constraints to permit effective design exploration. If determining feasibility required computationally expensive simulations, the cost of exploration would be prohibitive. Bayesian search is data-efficient...

Full description

Published in: Machine Learning, Optimization, and Data Science
ISBN: 9783030645793 9783030645809
ISSN: 0302-9743 1611-3349
Published: Cham Springer International Publishing 2021
Online Access: Check full text

URI: https://cronfa.swan.ac.uk/Record/cronfa55060
first_indexed 2020-08-25T08:25:29Z
last_indexed 2021-02-09T04:18:08Z
id cronfa55060
recordtype SURis
fullrecord <?xml version="1.0"?><rfc1807><datestamp>2021-02-08T17:00:12.1418923</datestamp><bib-version>v2</bib-version><id>55060</id><entry>2020-08-25</entry><title>On Bayesian Search for the Feasible Space Under Computationally Expensive Constraints</title><swanseaauthors><author><sid>6206f027aca1e3a5ff6b8cd224248bc2</sid><ORCID>0000-0002-5023-1371</ORCID><firstname>Alma</firstname><surname>Rahat</surname><name>Alma Rahat</name><active>true</active><ethesisStudent>false</ethesisStudent></author></swanseaauthors><date>2020-08-25</date><deptcode>MACS</deptcode><abstract>We are often interested in identifying the feasible subset of a decision space under multiple constraints to permit effective design exploration. If determining feasibility required computationally expensive simulations, the cost of exploration would be prohibitive. Bayesian search is data-efficient for such problems: starting from a small dataset, the central concept is to use Bayesian models of constraints with an acquisition function to locate promising solutions that may improve predictions of feasibility when the dataset is augmented. At the end of this sequential active learning approach with a limited number of expensive evaluations, the models can accurately predict the feasibility of any solution obviating the need for full simulations. In this paper, we propose a novel acquisition function that combines the probability that a solution lies at the boundary between feasible and infeasible spaces (representing exploitation) and the entropy in predictions (representing exploration). Experiments confirmed the efficacy of the proposed function.</abstract><type>Conference Paper/Proceeding/Abstract</type><journal>Machine Learning, Optimization, and Data Science</journal><volume>12566</volume><journalNumber/><paginationStart>529</paginationStart><paginationEnd>540</paginationEnd><publisher>Springer International Publishing</publisher><placeOfPublication>Cham</placeOfPublication><isbnPrint>9783030645793</isbnPrint><isbnElectronic>9783030645809</isbnElectronic><issnPrint>0302-9743</issnPrint><issnElectronic>1611-3349</issnElectronic><keywords>Active learning; Feasible region; Feasible design exploration; Gaussian processes; Constrained problems</keywords><publishedDay>7</publishedDay><publishedMonth>1</publishedMonth><publishedYear>2021</publishedYear><publishedDate>2021-01-07</publishedDate><doi>10.1007/978-3-030-64580-9_44</doi><url/><notes/><college>COLLEGE NANME</college><department>Mathematics and Computer Science School</department><CollegeCode>COLLEGE CODE</CollegeCode><DepartmentCode>MACS</DepartmentCode><institution>Swansea University</institution><apcterm/><lastEdited>2021-02-08T17:00:12.1418923</lastEdited><Created>2020-08-25T09:18:36.1990943</Created><path><level id="1">Faculty of Science and Engineering</level><level id="2">School of Mathematics and Computer Science - Computer Science</level></path><authors><author><firstname>Alma</firstname><surname>Rahat</surname><orcid>0000-0002-5023-1371</orcid><order>1</order></author><author><firstname>Michael</firstname><surname>Wood</surname><order>2</order></author></authors><documents><document><filename>55060__18033__f3b41226be76465c8fb171659d250945.pdf</filename><originalFilename>paper.pdf</originalFilename><uploaded>2020-08-25T09:23:44.5111315</uploaded><type>Output</type><contentLength>1200693</contentLength><contentType>application/pdf</contentType><version>Accepted Manuscript</version><cronfaStatus>true</cronfaStatus><copyrightCorrect>true</copyrightCorrect><language>eng</language></document></documents><OutputDurs/></rfc1807>
spelling 2021-02-08T17:00:12.1418923 v2 55060 2020-08-25 On Bayesian Search for the Feasible Space Under Computationally Expensive Constraints 6206f027aca1e3a5ff6b8cd224248bc2 0000-0002-5023-1371 Alma Rahat Alma Rahat true false 2020-08-25 MACS We are often interested in identifying the feasible subset of a decision space under multiple constraints to permit effective design exploration. If determining feasibility required computationally expensive simulations, the cost of exploration would be prohibitive. Bayesian search is data-efficient for such problems: starting from a small dataset, the central concept is to use Bayesian models of constraints with an acquisition function to locate promising solutions that may improve predictions of feasibility when the dataset is augmented. At the end of this sequential active learning approach with a limited number of expensive evaluations, the models can accurately predict the feasibility of any solution obviating the need for full simulations. In this paper, we propose a novel acquisition function that combines the probability that a solution lies at the boundary between feasible and infeasible spaces (representing exploitation) and the entropy in predictions (representing exploration). Experiments confirmed the efficacy of the proposed function. Conference Paper/Proceeding/Abstract Machine Learning, Optimization, and Data Science 12566 529 540 Springer International Publishing Cham 9783030645793 9783030645809 0302-9743 1611-3349 Active learning; Feasible region; Feasible design exploration; Gaussian processes; Constrained problems 7 1 2021 2021-01-07 10.1007/978-3-030-64580-9_44 COLLEGE NANME Mathematics and Computer Science School COLLEGE CODE MACS Swansea University 2021-02-08T17:00:12.1418923 2020-08-25T09:18:36.1990943 Faculty of Science and Engineering School of Mathematics and Computer Science - Computer Science Alma Rahat 0000-0002-5023-1371 1 Michael Wood 2 55060__18033__f3b41226be76465c8fb171659d250945.pdf paper.pdf 2020-08-25T09:23:44.5111315 Output 1200693 application/pdf Accepted Manuscript true true eng
title On Bayesian Search for the Feasible Space Under Computationally Expensive Constraints
spellingShingle On Bayesian Search for the Feasible Space Under Computationally Expensive Constraints
Alma Rahat
title_short On Bayesian Search for the Feasible Space Under Computationally Expensive Constraints
title_full On Bayesian Search for the Feasible Space Under Computationally Expensive Constraints
title_fullStr On Bayesian Search for the Feasible Space Under Computationally Expensive Constraints
title_full_unstemmed On Bayesian Search for the Feasible Space Under Computationally Expensive Constraints
title_sort On Bayesian Search for the Feasible Space Under Computationally Expensive Constraints
author_id_str_mv 6206f027aca1e3a5ff6b8cd224248bc2
author_id_fullname_str_mv 6206f027aca1e3a5ff6b8cd224248bc2_***_Alma Rahat
author Alma Rahat
author2 Alma Rahat
Michael Wood
format Conference Paper/Proceeding/Abstract
container_title Machine Learning, Optimization, and Data Science
container_volume 12566
container_start_page 529
publishDate 2021
institution Swansea University
isbn 9783030645793
9783030645809
issn 0302-9743
1611-3349
doi_str_mv 10.1007/978-3-030-64580-9_44
publisher Springer International Publishing
college_str Faculty of Science and Engineering
hierarchytype
hierarchy_top_id facultyofscienceandengineering
hierarchy_top_title Faculty of Science and Engineering
hierarchy_parent_id facultyofscienceandengineering
hierarchy_parent_title Faculty of Science and Engineering
department_str School of Mathematics and Computer Science - Computer Science{{{_:::_}}}Faculty of Science and Engineering{{{_:::_}}}School of Mathematics and Computer Science - Computer Science
document_store_str 1
active_str 0
description We are often interested in identifying the feasible subset of a decision space under multiple constraints to permit effective design exploration. If determining feasibility required computationally expensive simulations, the cost of exploration would be prohibitive. Bayesian search is data-efficient for such problems: starting from a small dataset, the central concept is to use Bayesian models of constraints with an acquisition function to locate promising solutions that may improve predictions of feasibility when the dataset is augmented. At the end of this sequential active learning approach with a limited number of expensive evaluations, the models can accurately predict the feasibility of any solution obviating the need for full simulations. In this paper, we propose a novel acquisition function that combines the probability that a solution lies at the boundary between feasible and infeasible spaces (representing exploitation) and the entropy in predictions (representing exploration). Experiments confirmed the efficacy of the proposed function.
published_date 2021-01-07T19:56:23Z
_version_ 1821346086034341888
score 11.04748

Similar Items