Neighborhood-based stopping criterion for contrastive divergence

Restricted Boltzmann Machines (RBMs) are general unsupervised learning devices to ascertain generative models of data distributions. RBMs are often trained using the Contrastive Divergence (CD) learning algorithm, an approximation to the gradient of the data log-likelihood (logL). A simple reconstru...

Descripción completa

Detalles Bibliográficos
Autores: Romero Merino, Enrique|||0000-0003-2404-5716, Mazzanti Castrillejo, Fernando Pablo|||0000-0001-6641-0609, Delgado Pin, Jordi|||0000-0003-4546-8355
Tipo de recurso: artículo
Fecha de publicación:2018
País:España
Institución:Universitat Politècnica de Catalunya (UPC)
Repositorio:UPCommons. Portal del coneixement obert de la UPC
Idioma:inglés
OAI Identifier:oai:upcommons.upc.edu:2117/125738
Acceso en línea:https://hdl.handle.net/2117/125738
https://dx.doi.org/10.1109/TNNLS.2017.2697455
Access Level:acceso abierto
Palabra clave:Machine learning
Neural networks (Computer science)
Recurrent neural networks
Restricted Boltzmann machines
Unsupervised learning
Aprenentatge automàtic
Xarxes neuronals (Informàtica)
Àrees temàtiques de la UPC::Informàtica::Intel·ligència artificial::Aprenentatge automàtic
id ES_2476dc4ce2c79b0abe4a9ba16ee8bb45
oai_identifier_str oai:upcommons.upc.edu:2117/125738
network_acronym_str ES
network_name_str España
repository_id_str
spelling Neighborhood-based stopping criterion for contrastive divergenceRomero Merino, Enrique|||0000-0003-2404-5716Mazzanti Castrillejo, Fernando Pablo|||0000-0001-6641-0609Delgado Pin, Jordi|||0000-0003-4546-8355Machine learningNeural networks (Computer science)Recurrent neural networksRestricted Boltzmann machinesUnsupervised learningAprenentatge automàticXarxes neuronals (Informàtica)Àrees temàtiques de la UPC::Informàtica::Intel·ligència artificial::Aprenentatge automàticRestricted Boltzmann Machines (RBMs) are general unsupervised learning devices to ascertain generative models of data distributions. RBMs are often trained using the Contrastive Divergence (CD) learning algorithm, an approximation to the gradient of the data log-likelihood (logL). A simple reconstruction error is often used as a stopping criterion for CD, although several authors have raised doubts concerning the feasibility of this procedure. In many cases, the evolution curve of the reconstruction error is monotonic, while the logL is not, thus indicating that the former is not a good estimator of the optimal stopping point for learning. However, not many alternatives to the reconstruction error have been discussed in the literature. An estimation of the logL of the training data based on annealed importance sampling is feasible but computationally very expensive. In this manuscript, we present a simple and cheap alternative, based on the inclusion of information contained in neighboring states to the training set, as a stopping criterion for CD learning.Peer Reviewed20182018-07-0120182018-12-13journal articlehttp://purl.org/coar/resource_type/c_6501AMhttp://purl.org/coar/version/c_ab4af688f83e57aainfo:eu-repo/semantics/articleapplication/pdfhttps://hdl.handle.net/2117/125738https://dx.doi.org/10.1109/TNNLS.2017.2697455reponame:UPCommons. Portal del coneixement obert de la UPCinstname:Universitat Politècnica de Catalunya (UPC)InglésengMinisterio de Economía y Competitividad http://doi.org/10.13039/501100003329 FIS2014-56257-C2-1-P MATERIA CUÁNTICA ULTRAFRIAMinisterio de Economía y Competitividad http://doi.org/10.13039/501100003329 TIN2014-57226-P APRENDIZAJE COMPUTACIONAL Y COMUNICACIONopen accesshttp://purl.org/coar/access_right/c_abf2info:eu-repo/semantics/openAccessoai:upcommons.upc.edu:2117/1257382026-05-27T15:37:01Z
dc.title.none.fl_str_mv Neighborhood-based stopping criterion for contrastive divergence
title Neighborhood-based stopping criterion for contrastive divergence
spellingShingle Neighborhood-based stopping criterion for contrastive divergence
Romero Merino, Enrique|||0000-0003-2404-5716
Machine learning
Neural networks (Computer science)
Recurrent neural networks
Restricted Boltzmann machines
Unsupervised learning
Aprenentatge automàtic
Xarxes neuronals (Informàtica)
Àrees temàtiques de la UPC::Informàtica::Intel·ligència artificial::Aprenentatge automàtic
title_short Neighborhood-based stopping criterion for contrastive divergence
title_full Neighborhood-based stopping criterion for contrastive divergence
title_fullStr Neighborhood-based stopping criterion for contrastive divergence
title_full_unstemmed Neighborhood-based stopping criterion for contrastive divergence
title_sort Neighborhood-based stopping criterion for contrastive divergence
dc.creator.none.fl_str_mv Romero Merino, Enrique|||0000-0003-2404-5716
Mazzanti Castrillejo, Fernando Pablo|||0000-0001-6641-0609
Delgado Pin, Jordi|||0000-0003-4546-8355
author Romero Merino, Enrique|||0000-0003-2404-5716
author_facet Romero Merino, Enrique|||0000-0003-2404-5716
Mazzanti Castrillejo, Fernando Pablo|||0000-0001-6641-0609
Delgado Pin, Jordi|||0000-0003-4546-8355
author_role author
author2 Mazzanti Castrillejo, Fernando Pablo|||0000-0001-6641-0609
Delgado Pin, Jordi|||0000-0003-4546-8355
author2_role author
author
dc.subject.none.fl_str_mv Machine learning
Neural networks (Computer science)
Recurrent neural networks
Restricted Boltzmann machines
Unsupervised learning
Aprenentatge automàtic
Xarxes neuronals (Informàtica)
Àrees temàtiques de la UPC::Informàtica::Intel·ligència artificial::Aprenentatge automàtic
topic Machine learning
Neural networks (Computer science)
Recurrent neural networks
Restricted Boltzmann machines
Unsupervised learning
Aprenentatge automàtic
Xarxes neuronals (Informàtica)
Àrees temàtiques de la UPC::Informàtica::Intel·ligència artificial::Aprenentatge automàtic
description Restricted Boltzmann Machines (RBMs) are general unsupervised learning devices to ascertain generative models of data distributions. RBMs are often trained using the Contrastive Divergence (CD) learning algorithm, an approximation to the gradient of the data log-likelihood (logL). A simple reconstruction error is often used as a stopping criterion for CD, although several authors have raised doubts concerning the feasibility of this procedure. In many cases, the evolution curve of the reconstruction error is monotonic, while the logL is not, thus indicating that the former is not a good estimator of the optimal stopping point for learning. However, not many alternatives to the reconstruction error have been discussed in the literature. An estimation of the logL of the training data based on annealed importance sampling is feasible but computationally very expensive. In this manuscript, we present a simple and cheap alternative, based on the inclusion of information contained in neighboring states to the training set, as a stopping criterion for CD learning.
publishDate 2018
dc.date.none.fl_str_mv 2018
2018-07-01
2018
2018-12-13
dc.type.none.fl_str_mv journal article
http://purl.org/coar/resource_type/c_6501
AM
http://purl.org/coar/version/c_ab4af688f83e57aa
dc.type.openaire.fl_str_mv info:eu-repo/semantics/article
format article
dc.identifier.none.fl_str_mv https://hdl.handle.net/2117/125738
https://dx.doi.org/10.1109/TNNLS.2017.2697455
url https://hdl.handle.net/2117/125738
https://dx.doi.org/10.1109/TNNLS.2017.2697455
dc.language.none.fl_str_mv Inglés
eng
language_invalid_str_mv Inglés
language eng
dc.relation.none.fl_str_mv Ministerio de Economía y Competitividad http://doi.org/10.13039/501100003329 FIS2014-56257-C2-1-P MATERIA CUÁNTICA ULTRAFRIA
Ministerio de Economía y Competitividad http://doi.org/10.13039/501100003329 TIN2014-57226-P APRENDIZAJE COMPUTACIONAL Y COMUNICACION
dc.rights.none.fl_str_mv open access
http://purl.org/coar/access_right/c_abf2
dc.rights.openaire.fl_str_mv info:eu-repo/semantics/openAccess
rights_invalid_str_mv open access
http://purl.org/coar/access_right/c_abf2
eu_rights_str_mv openAccess
dc.format.none.fl_str_mv application/pdf
dc.source.none.fl_str_mv reponame:UPCommons. Portal del coneixement obert de la UPC
instname:Universitat Politècnica de Catalunya (UPC)
instname_str Universitat Politècnica de Catalunya (UPC)
reponame_str UPCommons. Portal del coneixement obert de la UPC
collection UPCommons. Portal del coneixement obert de la UPC
repository.name.fl_str_mv
repository.mail.fl_str_mv
_version_ 1869404701126033408
score 15.301603

Similares en LA Referencia