Publicação

GSGP-CUDA: A CUDA framework for Geometric Semantic Genetic Programming

Ver documento

Detalhes bibliográficos
Resumo:Geometric Semantic Genetic Programming (GSGP) is a state-of-the-art machine learning method based on evolutionary computation. GSGP performs search operations directly at the level of program semantics, which can be done more efficiently than operating at the syntax level like most GP systems. Efficient implementations of GSGP in C++ exploit this fact, but not to its full potential. This paper presents GSGP-CUDA, the first CUDA implementation of GSGP and the most efficient, which exploits the intrinsic parallelism of GSGP using GPUs. Results show speedups greater than 1, 000× relative to the state-of-the-art sequential implementation, during the model training process. Additionally, our implementation allows the user to seamlessly make inferences over new data through the best evolved model, opening the possibility of using GSGP on Big Data problems.
Autores principais:Trujillo, Leonardo
Outros Autores:Muñoz Contreras, Jose Manuel; Hernandez, Daniel E.; Castelli, Mauro; Tapia, Juan J.
Assunto:Genetic Programming Geometric Semantic Genetic Programming CUDA GPU Software Computer Science Applications
Ano:2022
País:Portugal
Tipo de documento:artigo
Tipo de acesso:acesso aberto
Instituição associada:Universidade Nova de Lisboa
Idioma:inglês
Origem:Repositório Institucional da UNL
_version_ 1868983724115230720
author Trujillo, Leonardo
author2 Muñoz Contreras, Jose Manuel
Hernandez, Daniel E.
Castelli, Mauro
Tapia, Juan J.
author2_role author
author
author
author
author_facet Trujillo, Leonardo
Muñoz Contreras, Jose Manuel
Hernandez, Daniel E.
Castelli, Mauro
Tapia, Juan J.
author_role author
contributor_name_str_mv NOVA Information Management School (NOVA IMS)
Information Management Research Center (MagIC) - NOVA Information Management School
Elsevier BV
RUN
country_str PT
creators_json_txt [{\"Person.name\":\"Trujillo, Leonardo\"},{\"Person.name\":\"Muñoz Contreras, Jose Manuel\"},{\"Person.name\":\"Hernandez, Daniel E.\"},{\"Person.name\":\"Castelli, Mauro\"},{\"Person.name\":\"Tapia, Juan J.\"}]
datacite.contributors.contributor.contributorName.fl_str_mv NOVA Information Management School (NOVA IMS)
Information Management Research Center (MagIC) - NOVA Information Management School
Elsevier BV
RUN
datacite.creators.creator.creatorName.fl_str_mv Trujillo, Leonardo
Muñoz Contreras, Jose Manuel
Hernandez, Daniel E.
Castelli, Mauro
Tapia, Juan J.
datacite.date.Accepted.fl_str_mv 2022-06-01T00:00:00Z
datacite.date.available.fl_str_mv 2022-05-18T22:47:45Z
datacite.date.embargoed.fl_str_mv 2022-05-18T22:47:45Z
datacite.rights.fl_str_mv http://purl.org/coar/access_right/c_abf2
datacite.subjects.subject.fl_str_mv Genetic Programming
Geometric Semantic Genetic Programming
CUDA
GPU
Software
Computer Science Applications
datacite.titles.title.fl_str_mv GSGP-CUDA: A CUDA framework for Geometric Semantic Genetic Programming
dc.contributor.none.fl_str_mv NOVA Information Management School (NOVA IMS)
Information Management Research Center (MagIC) - NOVA Information Management School
Elsevier BV
RUN
dc.creator.none.fl_str_mv Trujillo, Leonardo
Muñoz Contreras, Jose Manuel
Hernandez, Daniel E.
Castelli, Mauro
Tapia, Juan J.
dc.date.Accepted.fl_str_mv 2022-06-01T00:00:00Z
dc.date.available.fl_str_mv 2022-05-18T22:47:45Z
dc.date.embargoed.fl_str_mv 2022-05-18T22:47:45Z
dc.format.none.fl_str_mv application/pdf
dc.identifier.none.fl_str_mv http://hdl.handle.net/10362/138207
dc.language.none.fl_str_mv eng
dc.rights.none.fl_str_mv http://purl.org/coar/access_right/c_abf2
dc.subject.none.fl_str_mv Genetic Programming
Geometric Semantic Genetic Programming
CUDA
GPU
Software
Computer Science Applications
dc.title.fl_str_mv GSGP-CUDA: A CUDA framework for Geometric Semantic Genetic Programming
dc.type.none.fl_str_mv http://purl.org/coar/resource_type/c_6501
description Geometric Semantic Genetic Programming (GSGP) is a state-of-the-art machine learning method based on evolutionary computation. GSGP performs search operations directly at the level of program semantics, which can be done more efficiently than operating at the syntax level like most GP systems. Efficient implementations of GSGP in C++ exploit this fact, but not to its full potential. This paper presents GSGP-CUDA, the first CUDA implementation of GSGP and the most efficient, which exploits the intrinsic parallelism of GSGP using GPUs. Results show speedups greater than 1, 000× relative to the state-of-the-art sequential implementation, during the model training process. Additionally, our implementation allows the user to seamlessly make inferences over new data through the best evolved model, opening the possibility of using GSGP on Big Data problems.
dirty 0
eu_rights_str_mv openAccess
format article
fulltext.url.fl_str_mv https://run.unl.pt/bitstreams/f03988cc-fc88-4847-8473-22b4a82c4b03/download
funder_facet_str_mv FCT{{{_:::_}}}Fundação para a Ciência e a Tecnologia
funding.funder.alternateName_str_mv FCT
funding.funder.identifier_str_mv http://doi.org/10.13039/501100001871
funding.funder.name_str_mv Fundação para a Ciência e a Tecnologia
funding.name_str_mv 3599-PPCDT
id run_d344ec33eeb201a24c19a4bb3f5d8578
identifier.url.fl_str_mv http://hdl.handle.net/10362/138207
inst_facet_str urn:organizationAcronym:unl{{{_:::_}}}Universidade Nova de Lisboa
instacron_str unl
institution Universidade Nova de Lisboa
instname_str Universidade Nova de Lisboa
language eng
network_acronym_str run
network_name_str Repositório Institucional da UNL
oai_identifier_str oai:run.unl.pt:10362/138207
organization_str_mv urn:organizationAcronym:unl
person_str_mv Trujillo, Leonardo
Muñoz Contreras, Jose Manuel
Hernandez, Daniel E.
Castelli, Mauro
Tapia, Juan J.
publishDate 2022
repo_facet_str urn:repositoryAcronym:run{{{_:::_}}}Repositório Institucional da UNL
reponame_str Repositório Institucional da UNL
repository_id_str urn:repositoryAcronym:run
service_str_mv urn:repositoryAcronym:run
spelling engenGeometric Semantic Genetic Programming (GSGP) is a state-of-the-art machine learning method based on evolutionary computation. GSGP performs search operations directly at the level of program semantics, which can be done more efficiently than operating at the syntax level like most GP systems. Efficient implementations of GSGP in C++ exploit this fact, but not to its full potential. This paper presents GSGP-CUDA, the first CUDA implementation of GSGP and the most efficient, which exploits the intrinsic parallelism of GSGP using GPUs. Results show speedups greater than 1, 000× relative to the state-of-the-art sequential implementation, during the model training process. Additionally, our implementation allows the user to seamlessly make inferences over new data through the best evolved model, opening the possibility of using GSGP on Big Data problems.application/pdfenGSGP-CUDA: A CUDA framework for Geometric Semantic Genetic ProgrammingTrujillo, LeonardoMuñoz Contreras, Jose ManuelHernandez, Daniel E.Castelli, MauroTapia, Juan J.NOVA Information Management School (NOVA IMS)Information Management Research Center (MagIC) - NOVA Information Management SchoolElsevier BVHostingInstitutionOrganizationalRUNe-mailmailto:run@unl.ptrun@unl.ptURNIsPartOfPURE: 43986256URNIsPartOfPURE UUID: f45ab250-dc7f-46fe-8049-600e5a6b0f7aURNIsPartOfcrossref: 10.1016/j.softx.2022.101085URNIsPartOfScopus: 85129991202URNIsPartOfWOS: 000800216100003URNIsPartOfORCID: /0000-0002-8793-1451/work/113263202DOIIsPartOf10.1016/j.softx.2022.1010852022-05-18T22:47:45Z2022-06-012022-06-01T00:00:00ZHandlehttp://hdl.handle.net/10362/138207http://purl.org/coar/access_right/c_abf2open accessGenetic ProgrammingGeometric Semantic Genetic ProgrammingCUDAGPUSoftwareComputer Science Applications380112 bytesFundação para a Ciência e a TecnologiaData Science and Over-Indebtedness: Use of Artificial Intelligence Algorithms in Credit Consumption and Indebtedness Conciliation in Portugal3599-PPCDTCrossref Funder IDhttp://doi.org/10.13039/501100001871literaturehttp://purl.org/coar/resource_type/c_6501journal articlehttp://purl.org/coar/access_right/c_abf2application/pdffulltexthttps://run.unl.pt/bitstreams/f03988cc-fc88-4847-8473-22b4a82c4b03/download
spellingShingle GSGP-CUDA: A CUDA framework for Geometric Semantic Genetic Programming
Trujillo, Leonardo
Genetic Programming
Geometric Semantic Genetic Programming
CUDA
GPU
Software
Computer Science Applications
status SINGLETON
subject.fl_str_mv Genetic Programming
Geometric Semantic Genetic Programming
CUDA
GPU
Software
Computer Science Applications
title GSGP-CUDA: A CUDA framework for Geometric Semantic Genetic Programming
title_full GSGP-CUDA: A CUDA framework for Geometric Semantic Genetic Programming
title_fullStr GSGP-CUDA: A CUDA framework for Geometric Semantic Genetic Programming
title_full_unstemmed GSGP-CUDA: A CUDA framework for Geometric Semantic Genetic Programming
title_short GSGP-CUDA: A CUDA framework for Geometric Semantic Genetic Programming
title_sort GSGP-CUDA: A CUDA framework for Geometric Semantic Genetic Programming
topic Genetic Programming
Geometric Semantic Genetic Programming
CUDA
GPU
Software
Computer Science Applications
topic_facet Genetic Programming
Geometric Semantic Genetic Programming
CUDA
GPU
Software
Computer Science Applications
url http://hdl.handle.net/10362/138207
visible 1