Deep Boltzmann machines using adaptive temperatures
Autor(a) principal: | |
---|---|
Data de Publicação: | 2017 |
Outros Autores: | , |
Tipo de documento: | Artigo de conferência |
Idioma: | eng |
Título da fonte: | Repositório Institucional da UNESP |
Texto Completo: | http://dx.doi.org/10.1007/978-3-319-64689-3_14 http://hdl.handle.net/11449/179135 |
Resumo: | Deep learning has been considered a hallmark in a number of applications recently. Among those techniques, the ones based on Restricted Boltzmann Machines have attracted a considerable attention, since they are energy-driven models composed of latent variables that aim at learning the probability distribution of the input data. In a nutshell, the training procedure of such models concerns the minimization of the energy of each training sample in order to increase its probability. Therefore, such optimization process needs to be regularized in order to reach the best trade-off between exploitation and exploration. In this work, we propose an adaptive regularization approach based on temperatures, and we show its advantages considering Deep Belief Networks (DBNs) and Deep Boltzmann Machines (DBMs). The proposed approach is evaluated in the context of binary image reconstruction, thus outperforming temperature-fixed DBNs and DBMs. |
id |
UNSP_2256b63861663b5f217027d186c5b94e |
---|---|
oai_identifier_str |
oai:repositorio.unesp.br:11449/179135 |
network_acronym_str |
UNSP |
network_name_str |
Repositório Institucional da UNESP |
repository_id_str |
2946 |
spelling |
Deep Boltzmann machines using adaptive temperaturesDeep learning has been considered a hallmark in a number of applications recently. Among those techniques, the ones based on Restricted Boltzmann Machines have attracted a considerable attention, since they are energy-driven models composed of latent variables that aim at learning the probability distribution of the input data. In a nutshell, the training procedure of such models concerns the minimization of the energy of each training sample in order to increase its probability. Therefore, such optimization process needs to be regularized in order to reach the best trade-off between exploitation and exploration. In this work, we propose an adaptive regularization approach based on temperatures, and we show its advantages considering Deep Belief Networks (DBNs) and Deep Boltzmann Machines (DBMs). The proposed approach is evaluated in the context of binary image reconstruction, thus outperforming temperature-fixed DBNs and DBMs.Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)Department of Computing UFSCar - Federal University of São CarlosSchool of Sciences UNESP - São Paulo State UniversitySchool of Sciences UNESP - São Paulo State UniversityFAPESP: #2014/12236-1FAPESP: #2014/16250-9CNPq: #306166/2014-3Universidade Federal de São Carlos (UFSCar)Universidade Estadual Paulista (Unesp)Passos Júnior, Leandro A.Costa, Kelton A. P. [UNESP]Papa, João P. [UNESP]2018-12-11T17:33:53Z2018-12-11T17:33:53Z2017-01-01info:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/conferenceObject172-183http://dx.doi.org/10.1007/978-3-319-64689-3_14Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), v. 10424 LNCS, p. 172-183.1611-33490302-9743http://hdl.handle.net/11449/17913510.1007/978-3-319-64689-3_142-s2.0-85028518481Scopusreponame:Repositório Institucional da UNESPinstname:Universidade Estadual Paulista (UNESP)instacron:UNESPengLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)0,295info:eu-repo/semantics/openAccess2024-04-23T16:11:20Zoai:repositorio.unesp.br:11449/179135Repositório InstitucionalPUBhttp://repositorio.unesp.br/oai/requestopendoar:29462024-08-05T17:54:47.363134Repositório Institucional da UNESP - Universidade Estadual Paulista (UNESP)false |
dc.title.none.fl_str_mv |
Deep Boltzmann machines using adaptive temperatures |
title |
Deep Boltzmann machines using adaptive temperatures |
spellingShingle |
Deep Boltzmann machines using adaptive temperatures Passos Júnior, Leandro A. |
title_short |
Deep Boltzmann machines using adaptive temperatures |
title_full |
Deep Boltzmann machines using adaptive temperatures |
title_fullStr |
Deep Boltzmann machines using adaptive temperatures |
title_full_unstemmed |
Deep Boltzmann machines using adaptive temperatures |
title_sort |
Deep Boltzmann machines using adaptive temperatures |
author |
Passos Júnior, Leandro A. |
author_facet |
Passos Júnior, Leandro A. Costa, Kelton A. P. [UNESP] Papa, João P. [UNESP] |
author_role |
author |
author2 |
Costa, Kelton A. P. [UNESP] Papa, João P. [UNESP] |
author2_role |
author author |
dc.contributor.none.fl_str_mv |
Universidade Federal de São Carlos (UFSCar) Universidade Estadual Paulista (Unesp) |
dc.contributor.author.fl_str_mv |
Passos Júnior, Leandro A. Costa, Kelton A. P. [UNESP] Papa, João P. [UNESP] |
description |
Deep learning has been considered a hallmark in a number of applications recently. Among those techniques, the ones based on Restricted Boltzmann Machines have attracted a considerable attention, since they are energy-driven models composed of latent variables that aim at learning the probability distribution of the input data. In a nutshell, the training procedure of such models concerns the minimization of the energy of each training sample in order to increase its probability. Therefore, such optimization process needs to be regularized in order to reach the best trade-off between exploitation and exploration. In this work, we propose an adaptive regularization approach based on temperatures, and we show its advantages considering Deep Belief Networks (DBNs) and Deep Boltzmann Machines (DBMs). The proposed approach is evaluated in the context of binary image reconstruction, thus outperforming temperature-fixed DBNs and DBMs. |
publishDate |
2017 |
dc.date.none.fl_str_mv |
2017-01-01 2018-12-11T17:33:53Z 2018-12-11T17:33:53Z |
dc.type.status.fl_str_mv |
info:eu-repo/semantics/publishedVersion |
dc.type.driver.fl_str_mv |
info:eu-repo/semantics/conferenceObject |
format |
conferenceObject |
status_str |
publishedVersion |
dc.identifier.uri.fl_str_mv |
http://dx.doi.org/10.1007/978-3-319-64689-3_14 Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), v. 10424 LNCS, p. 172-183. 1611-3349 0302-9743 http://hdl.handle.net/11449/179135 10.1007/978-3-319-64689-3_14 2-s2.0-85028518481 |
url |
http://dx.doi.org/10.1007/978-3-319-64689-3_14 http://hdl.handle.net/11449/179135 |
identifier_str_mv |
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), v. 10424 LNCS, p. 172-183. 1611-3349 0302-9743 10.1007/978-3-319-64689-3_14 2-s2.0-85028518481 |
dc.language.iso.fl_str_mv |
eng |
language |
eng |
dc.relation.none.fl_str_mv |
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) 0,295 |
dc.rights.driver.fl_str_mv |
info:eu-repo/semantics/openAccess |
eu_rights_str_mv |
openAccess |
dc.format.none.fl_str_mv |
172-183 |
dc.source.none.fl_str_mv |
Scopus reponame:Repositório Institucional da UNESP instname:Universidade Estadual Paulista (UNESP) instacron:UNESP |
instname_str |
Universidade Estadual Paulista (UNESP) |
instacron_str |
UNESP |
institution |
UNESP |
reponame_str |
Repositório Institucional da UNESP |
collection |
Repositório Institucional da UNESP |
repository.name.fl_str_mv |
Repositório Institucional da UNESP - Universidade Estadual Paulista (UNESP) |
repository.mail.fl_str_mv |
|
_version_ |
1808128874284843008 |