Categorization in fully connected multistate neural network models
Autor(a) principal: | |
---|---|
Data de Publicação: | 1999 |
Outros Autores: | , |
Tipo de documento: | Artigo |
Idioma: | eng |
Título da fonte: | Repositório Institucional da UFRGS |
Texto Completo: | http://hdl.handle.net/10183/101307 |
Resumo: | The categorization ability of fully connected neural network models, with either discrete or continuous Q-state units, is studied in this work in replica symmetric mean-field theory. Hierarchically correlated multistate patterns in a two level structure of ancestors and descendents ~examples! are embedded in the network and the categorization task consists in recognizing the ancestors when the network is trained exclusively with their descendents. Explicit results for the dependence of the equilibrium properties of a Q=3-state model and a Q=∞ state model are obtained in the form of phase diagrams and categorization curves. A strong improvement of the categorization ability is found when the network is trained with examples of low activity. The categorization ability is found to be robust to finite threshold and synaptic noise. The Almeida-Thouless lines that limit the validity of the replica-symmetric results, are also obtained. [S1063-651X(99)09212-0] |
id |
UFRGS-2_568156555aacc1b8b458ab784b445f9b |
---|---|
oai_identifier_str |
oai:www.lume.ufrgs.br:10183/101307 |
network_acronym_str |
UFRGS-2 |
network_name_str |
Repositório Institucional da UFRGS |
repository_id_str |
|
spelling |
Erichsen Junior, RubemTheumann, Walter KarlDominguez, David Renato Carreta2014-08-19T02:10:10Z19991063-651Xhttp://hdl.handle.net/10183/101307000055631The categorization ability of fully connected neural network models, with either discrete or continuous Q-state units, is studied in this work in replica symmetric mean-field theory. Hierarchically correlated multistate patterns in a two level structure of ancestors and descendents ~examples! are embedded in the network and the categorization task consists in recognizing the ancestors when the network is trained exclusively with their descendents. Explicit results for the dependence of the equilibrium properties of a Q=3-state model and a Q=∞ state model are obtained in the form of phase diagrams and categorization curves. A strong improvement of the categorization ability is found when the network is trained with examples of low activity. The categorization ability is found to be robust to finite threshold and synaptic noise. The Almeida-Thouless lines that limit the validity of the replica-symmetric results, are also obtained. [S1063-651X(99)09212-0]application/pdfengPhysical Review. E, Statistical Physics, Plasmas, Fluids and Related Interdisciplinary Topics. New York. Vol. 60, no. 6, pt. B (Dec. 1999), p. 7321-7331Redes neurais de hopfieldRuídosCategorization in fully connected multistate neural network modelsEstrangeiroinfo:eu-repo/semantics/articleinfo:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/openAccessreponame:Repositório Institucional da UFRGSinstname:Universidade Federal do Rio Grande do Sul (UFRGS)instacron:UFRGSORIGINAL000055631.pdf000055631.pdfTexto completo (inglês)application/pdf123427http://www.lume.ufrgs.br/bitstream/10183/101307/1/000055631.pdf3e029a2362b593b0fb6eddf1d9e6c94aMD51TEXT000055631.pdf.txt000055631.pdf.txtExtracted Texttext/plain39094http://www.lume.ufrgs.br/bitstream/10183/101307/2/000055631.pdf.txt5c0c113a1aa6619a7d69758283c37765MD52THUMBNAIL000055631.pdf.jpg000055631.pdf.jpgGenerated Thumbnailimage/jpeg2006http://www.lume.ufrgs.br/bitstream/10183/101307/3/000055631.pdf.jpg409b37f683293972943dafad661f424aMD5310183/1013072018-10-22 09:19:42.323oai:www.lume.ufrgs.br:10183/101307Repositório de PublicaçõesPUBhttps://lume.ufrgs.br/oai/requestopendoar:2018-10-22T12:19:42Repositório Institucional da UFRGS - Universidade Federal do Rio Grande do Sul (UFRGS)false |
dc.title.pt_BR.fl_str_mv |
Categorization in fully connected multistate neural network models |
title |
Categorization in fully connected multistate neural network models |
spellingShingle |
Categorization in fully connected multistate neural network models Erichsen Junior, Rubem Redes neurais de hopfield Ruídos |
title_short |
Categorization in fully connected multistate neural network models |
title_full |
Categorization in fully connected multistate neural network models |
title_fullStr |
Categorization in fully connected multistate neural network models |
title_full_unstemmed |
Categorization in fully connected multistate neural network models |
title_sort |
Categorization in fully connected multistate neural network models |
author |
Erichsen Junior, Rubem |
author_facet |
Erichsen Junior, Rubem Theumann, Walter Karl Dominguez, David Renato Carreta |
author_role |
author |
author2 |
Theumann, Walter Karl Dominguez, David Renato Carreta |
author2_role |
author author |
dc.contributor.author.fl_str_mv |
Erichsen Junior, Rubem Theumann, Walter Karl Dominguez, David Renato Carreta |
dc.subject.por.fl_str_mv |
Redes neurais de hopfield Ruídos |
topic |
Redes neurais de hopfield Ruídos |
description |
The categorization ability of fully connected neural network models, with either discrete or continuous Q-state units, is studied in this work in replica symmetric mean-field theory. Hierarchically correlated multistate patterns in a two level structure of ancestors and descendents ~examples! are embedded in the network and the categorization task consists in recognizing the ancestors when the network is trained exclusively with their descendents. Explicit results for the dependence of the equilibrium properties of a Q=3-state model and a Q=∞ state model are obtained in the form of phase diagrams and categorization curves. A strong improvement of the categorization ability is found when the network is trained with examples of low activity. The categorization ability is found to be robust to finite threshold and synaptic noise. The Almeida-Thouless lines that limit the validity of the replica-symmetric results, are also obtained. [S1063-651X(99)09212-0] |
publishDate |
1999 |
dc.date.issued.fl_str_mv |
1999 |
dc.date.accessioned.fl_str_mv |
2014-08-19T02:10:10Z |
dc.type.driver.fl_str_mv |
Estrangeiro info:eu-repo/semantics/article |
dc.type.status.fl_str_mv |
info:eu-repo/semantics/publishedVersion |
format |
article |
status_str |
publishedVersion |
dc.identifier.uri.fl_str_mv |
http://hdl.handle.net/10183/101307 |
dc.identifier.issn.pt_BR.fl_str_mv |
1063-651X |
dc.identifier.nrb.pt_BR.fl_str_mv |
000055631 |
identifier_str_mv |
1063-651X 000055631 |
url |
http://hdl.handle.net/10183/101307 |
dc.language.iso.fl_str_mv |
eng |
language |
eng |
dc.relation.ispartof.pt_BR.fl_str_mv |
Physical Review. E, Statistical Physics, Plasmas, Fluids and Related Interdisciplinary Topics. New York. Vol. 60, no. 6, pt. B (Dec. 1999), p. 7321-7331 |
dc.rights.driver.fl_str_mv |
info:eu-repo/semantics/openAccess |
eu_rights_str_mv |
openAccess |
dc.format.none.fl_str_mv |
application/pdf |
dc.source.none.fl_str_mv |
reponame:Repositório Institucional da UFRGS instname:Universidade Federal do Rio Grande do Sul (UFRGS) instacron:UFRGS |
instname_str |
Universidade Federal do Rio Grande do Sul (UFRGS) |
instacron_str |
UFRGS |
institution |
UFRGS |
reponame_str |
Repositório Institucional da UFRGS |
collection |
Repositório Institucional da UFRGS |
bitstream.url.fl_str_mv |
http://www.lume.ufrgs.br/bitstream/10183/101307/1/000055631.pdf http://www.lume.ufrgs.br/bitstream/10183/101307/2/000055631.pdf.txt http://www.lume.ufrgs.br/bitstream/10183/101307/3/000055631.pdf.jpg |
bitstream.checksum.fl_str_mv |
3e029a2362b593b0fb6eddf1d9e6c94a 5c0c113a1aa6619a7d69758283c37765 409b37f683293972943dafad661f424a |
bitstream.checksumAlgorithm.fl_str_mv |
MD5 MD5 MD5 |
repository.name.fl_str_mv |
Repositório Institucional da UFRGS - Universidade Federal do Rio Grande do Sul (UFRGS) |
repository.mail.fl_str_mv |
|
_version_ |
1801224845279625216 |