PyTorch Genesis - A Framework for Neuroevolution Neurevolution Using Genetic Programming and Backpropagation

Detalhes bibliográficos
Autor(a) principal: Santos, Frederico José Jácome de Brito
Data de Publicação: 2023
Tipo de documento: Dissertação
Idioma: eng
Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
Texto Completo: http://hdl.handle.net/10362/148551
Resumo: Dissertation presented as the partial requirement for obtaining a Master's degree in Data Science and Advanced Analytics, specialization in Data Science
id RCAP_66008d64abfbaf56a4fbcb7d007f81a4
oai_identifier_str oai:run.unl.pt:10362/148551
network_acronym_str RCAP
network_name_str Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
repository_id_str 7160
spelling PyTorch Genesis - A Framework for Neuroevolution Neurevolution Using Genetic Programming and BackpropagationNeuroevolutionEvolutionary Deep LearningNeural Architecture SearchSupervised LearningDissertation presented as the partial requirement for obtaining a Master's degree in Data Science and Advanced Analytics, specialization in Data ScienceSelf-evolving neural networks have been long sought after. Evolutionary Deep Learning is an emerging field full of exciting research bringing together the fields of Deep Learning and Evolutionary Computation. The objective of this thesis is to develop a method of evolving deep neural networks by adapting the theory behind Geometric Semantic Genetic Programming, a subfield of Genetic Programming, and Semantic Learning Machine. Our method evolves neural networks by incrementing their number of neurons throughout generations, whilst using backpropagation for the optimization of the network’s parameters. We bring together evolution through natural selection and the advances in optimization through backpropagation in the field of deep learning. We evolve neural networks that achieve nearly 90% accuracy on the CIFAR-10 dataset with a relatively low number of parameters, evolving in GPU-minutes vs the field standard of GPU-days. We develop PyTorch Genesis, a framework to evolve these models with the hope of opening the gates to a different way of evolving neural networks.Castelli, MauroGonçalves, Ivo Carlos PereiraRUNSantos, Frederico José Jácome de Brito2023-01-232026-01-23T00:00:00Z2023-01-23T00:00:00Zinfo:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/masterThesisapplication/pdfhttp://hdl.handle.net/10362/148551TID:203210387enginfo:eu-repo/semantics/embargoedAccessreponame:Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)instname:Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informaçãoinstacron:RCAAP2024-03-11T05:30:13Zoai:run.unl.pt:10362/148551Portal AgregadorONGhttps://www.rcaap.pt/oai/openaireopendoar:71602024-03-20T03:53:25.056658Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) - Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informaçãofalse
dc.title.none.fl_str_mv PyTorch Genesis - A Framework for Neuroevolution Neurevolution Using Genetic Programming and Backpropagation
title PyTorch Genesis - A Framework for Neuroevolution Neurevolution Using Genetic Programming and Backpropagation
spellingShingle PyTorch Genesis - A Framework for Neuroevolution Neurevolution Using Genetic Programming and Backpropagation
Santos, Frederico José Jácome de Brito
Neuroevolution
Evolutionary Deep Learning
Neural Architecture Search
Supervised Learning
title_short PyTorch Genesis - A Framework for Neuroevolution Neurevolution Using Genetic Programming and Backpropagation
title_full PyTorch Genesis - A Framework for Neuroevolution Neurevolution Using Genetic Programming and Backpropagation
title_fullStr PyTorch Genesis - A Framework for Neuroevolution Neurevolution Using Genetic Programming and Backpropagation
title_full_unstemmed PyTorch Genesis - A Framework for Neuroevolution Neurevolution Using Genetic Programming and Backpropagation
title_sort PyTorch Genesis - A Framework for Neuroevolution Neurevolution Using Genetic Programming and Backpropagation
author Santos, Frederico José Jácome de Brito
author_facet Santos, Frederico José Jácome de Brito
author_role author
dc.contributor.none.fl_str_mv Castelli, Mauro
Gonçalves, Ivo Carlos Pereira
RUN
dc.contributor.author.fl_str_mv Santos, Frederico José Jácome de Brito
dc.subject.por.fl_str_mv Neuroevolution
Evolutionary Deep Learning
Neural Architecture Search
Supervised Learning
topic Neuroevolution
Evolutionary Deep Learning
Neural Architecture Search
Supervised Learning
description Dissertation presented as the partial requirement for obtaining a Master's degree in Data Science and Advanced Analytics, specialization in Data Science
publishDate 2023
dc.date.none.fl_str_mv 2023-01-23
2023-01-23T00:00:00Z
2026-01-23T00:00:00Z
dc.type.status.fl_str_mv info:eu-repo/semantics/publishedVersion
dc.type.driver.fl_str_mv info:eu-repo/semantics/masterThesis
format masterThesis
status_str publishedVersion
dc.identifier.uri.fl_str_mv http://hdl.handle.net/10362/148551
TID:203210387
url http://hdl.handle.net/10362/148551
identifier_str_mv TID:203210387
dc.language.iso.fl_str_mv eng
language eng
dc.rights.driver.fl_str_mv info:eu-repo/semantics/embargoedAccess
eu_rights_str_mv embargoedAccess
dc.format.none.fl_str_mv application/pdf
dc.source.none.fl_str_mv reponame:Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
instname:Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação
instacron:RCAAP
instname_str Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação
instacron_str RCAAP
institution RCAAP
reponame_str Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
collection Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
repository.name.fl_str_mv Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) - Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação
repository.mail.fl_str_mv
_version_ 1799138124592316416

Registros relacionados