NEURON NETWORKS
Ciência E Natura
Endereço:
Revista Ciência e Natura |
Campus Sede-Cidade Universitária |
Av. Roraima nº 1000, Prédio 13, Sala 1122 |
Fone/Fax +55(55) 3220-8735 |
Bairro Camobi
Santa Maria / RS
97105-900
Site: http://www.ufsm.br/cienciaenatura
Telefone: (55) 3220-8735
ISSN: 2179-460X
Editor Chefe: Marcelo Barcellos da Rosa
Início Publicação: 30/11/1979
Periodicidade: Quadrimestral
NEURON NETWORKS
Ano: 1992 | Volume: 14 | Número: 14
Autores: Rita MC de Almeida
Autor Correspondente: Rita MC de Almeida | [email protected]
Autor Correspondente: Rita MC de Almeida | [email protected]
Palavras-chave: there are no notes
Resumos Cadastrados
Resumo Inglês:
In the last decade several discoveries in relation to neurons and how they are interconnected, forming networks, made possible the study of the dynamics of the storage and processing of information by the brain. In particular, the physicist JJHopfield proposed a formal, minimalist model for these neural networks, reducing the problem to a particular case of a well-defined physical system - spin glass. Although the problem is well defined, its solution is far from trivial.
In this text we introduce the problem, describe the Hopfield model with its results and limitations and present our contribution to the description of the information storage in networks of neurons