Space of interactions with definite symmetry in neural networks with biased patterns as a spin-glass problem
Autor(a) principal: | |
---|---|
Data de Publicação: | 1996 |
Tipo de documento: | Artigo |
Idioma: | eng |
Título da fonte: | Repositório Institucional da UFRGS |
Texto Completo: | http://hdl.handle.net/10183/103640 |
Resumo: | We study the space of interactions of a connected neural network with biased patterns, when the synaptic interactions satisfy a symmetry constraint. We show that the solution to the problem requires the calculation of a quantity NΩμ analogous to the thermodynamic potential of a multiply connected Ising model with site dependent interactions, which maps the present problem into the spin-glass problem. By using a diagrammatic expansion, we express NΩμ formally as a functional of renormalized site dependent ‘‘propagators’’ Gij and local ‘‘magnetizations’’ mi , which are determined from a variational principle. Calculating NΩμ in the single site or Brout approximation we recover the theory of Thouless, Anderson, and Palmer (TAP), while the mi satisfy TAP-like equations. In the impossibility of solving the equations, we analyze an approximate solution that sums only tree diagrams and interpolates between the two known results of total asymmetry, finite bias, and arbitrary symmetry with vanishing bias. The results show a small dependence on the asymmetry parameter. |
id |
UFRGS-2_e2bebfd3f24edae6b9120df95543d6eb |
---|---|
oai_identifier_str |
oai:www.lume.ufrgs.br:10183/103640 |
network_acronym_str |
UFRGS-2 |
network_name_str |
Repositório Institucional da UFRGS |
repository_id_str |
|
spelling |
Theumann, Alba Graciela Rivas de2014-09-23T02:12:32Z19961063-651Xhttp://hdl.handle.net/10183/103640000148973We study the space of interactions of a connected neural network with biased patterns, when the synaptic interactions satisfy a symmetry constraint. We show that the solution to the problem requires the calculation of a quantity NΩμ analogous to the thermodynamic potential of a multiply connected Ising model with site dependent interactions, which maps the present problem into the spin-glass problem. By using a diagrammatic expansion, we express NΩμ formally as a functional of renormalized site dependent ‘‘propagators’’ Gij and local ‘‘magnetizations’’ mi , which are determined from a variational principle. Calculating NΩμ in the single site or Brout approximation we recover the theory of Thouless, Anderson, and Palmer (TAP), while the mi satisfy TAP-like equations. In the impossibility of solving the equations, we analyze an approximate solution that sums only tree diagrams and interpolates between the two known results of total asymmetry, finite bias, and arbitrary symmetry with vanishing bias. The results show a small dependence on the asymmetry parameter.application/pdfengPhysical Review. E, Statistical physics, plasmas, fluids and related interdisciplinary topics. New York. Vol. 53, no. 6B (June 1996), p. 6361-6370Física da matéria condensadaRedes neuraisSimetriasVidros de spinModelo de isingMagnetizaçãoRenormalizacaoModelos de vidros de spinBiofísicaMatemáticaMecânica estatísticaEquilíbrio de faseSpace of interactions with definite symmetry in neural networks with biased patterns as a spin-glass problemEstrangeiroinfo: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:UFRGSORIGINAL000148973.pdf000148973.pdfTexto completo (inglês)application/pdf162627http://www.lume.ufrgs.br/bitstream/10183/103640/1/000148973.pdf43ea5012f8d384272f686e22945aa49aMD51TEXT000148973.pdf.txt000148973.pdf.txtExtracted Texttext/plain28981http://www.lume.ufrgs.br/bitstream/10183/103640/2/000148973.pdf.txt07898cbbb87bd8d9c7b24d1997dd79dcMD52THUMBNAIL000148973.pdf.jpg000148973.pdf.jpgGenerated Thumbnailimage/jpeg2048http://www.lume.ufrgs.br/bitstream/10183/103640/3/000148973.pdf.jpgd4f0d0ed6edc1166b4bfae42e831dbb0MD5310183/1036402018-10-10 08:08:59.151oai:www.lume.ufrgs.br:10183/103640Repositório de PublicaçõesPUBhttps://lume.ufrgs.br/oai/requestopendoar:2018-10-10T11:08:59Repositório Institucional da UFRGS - Universidade Federal do Rio Grande do Sul (UFRGS)false |
dc.title.pt_BR.fl_str_mv |
Space of interactions with definite symmetry in neural networks with biased patterns as a spin-glass problem |
title |
Space of interactions with definite symmetry in neural networks with biased patterns as a spin-glass problem |
spellingShingle |
Space of interactions with definite symmetry in neural networks with biased patterns as a spin-glass problem Theumann, Alba Graciela Rivas de Física da matéria condensada Redes neurais Simetrias Vidros de spin Modelo de ising Magnetização Renormalizacao Modelos de vidros de spin Biofísica Matemática Mecânica estatística Equilíbrio de fase |
title_short |
Space of interactions with definite symmetry in neural networks with biased patterns as a spin-glass problem |
title_full |
Space of interactions with definite symmetry in neural networks with biased patterns as a spin-glass problem |
title_fullStr |
Space of interactions with definite symmetry in neural networks with biased patterns as a spin-glass problem |
title_full_unstemmed |
Space of interactions with definite symmetry in neural networks with biased patterns as a spin-glass problem |
title_sort |
Space of interactions with definite symmetry in neural networks with biased patterns as a spin-glass problem |
author |
Theumann, Alba Graciela Rivas de |
author_facet |
Theumann, Alba Graciela Rivas de |
author_role |
author |
dc.contributor.author.fl_str_mv |
Theumann, Alba Graciela Rivas de |
dc.subject.por.fl_str_mv |
Física da matéria condensada Redes neurais Simetrias Vidros de spin Modelo de ising Magnetização Renormalizacao Modelos de vidros de spin Biofísica Matemática Mecânica estatística Equilíbrio de fase |
topic |
Física da matéria condensada Redes neurais Simetrias Vidros de spin Modelo de ising Magnetização Renormalizacao Modelos de vidros de spin Biofísica Matemática Mecânica estatística Equilíbrio de fase |
description |
We study the space of interactions of a connected neural network with biased patterns, when the synaptic interactions satisfy a symmetry constraint. We show that the solution to the problem requires the calculation of a quantity NΩμ analogous to the thermodynamic potential of a multiply connected Ising model with site dependent interactions, which maps the present problem into the spin-glass problem. By using a diagrammatic expansion, we express NΩμ formally as a functional of renormalized site dependent ‘‘propagators’’ Gij and local ‘‘magnetizations’’ mi , which are determined from a variational principle. Calculating NΩμ in the single site or Brout approximation we recover the theory of Thouless, Anderson, and Palmer (TAP), while the mi satisfy TAP-like equations. In the impossibility of solving the equations, we analyze an approximate solution that sums only tree diagrams and interpolates between the two known results of total asymmetry, finite bias, and arbitrary symmetry with vanishing bias. The results show a small dependence on the asymmetry parameter. |
publishDate |
1996 |
dc.date.issued.fl_str_mv |
1996 |
dc.date.accessioned.fl_str_mv |
2014-09-23T02:12:32Z |
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/10183/103640 |
dc.identifier.issn.pt_BR.fl_str_mv |
1063-651X |
dc.identifier.nrb.pt_BR.fl_str_mv |
000148973 |
identifier_str_mv |
1063-651X 000148973 |
url |
http://hdl.handle.net/10183/103640 |
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. 53, no. 6B (June 1996), p. 6361-6370 |
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/10183/103640/1/000148973.pdf http://www.lume.ufrgs.br/bitstream/10183/103640/2/000148973.pdf.txt http://www.lume.ufrgs.br/bitstream/10183/103640/3/000148973.pdf.jpg |
bitstream.checksum.fl_str_mv |
43ea5012f8d384272f686e22945aa49a 07898cbbb87bd8d9c7b24d1997dd79dc d4f0d0ed6edc1166b4bfae42e831dbb0 |
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_ |
1801224850995412992 |