O modelo RS para redes de neurônios

View/ Open
Date
1991Author
Advisor
Co-advisor
Academic level
Master
Type
Abstract in Portuguese (Brasil)
Abstract
A model for neural networks is studied through a multineuron interaction energy function E given by the product of the squared distances in phase space between the state S of the net and the P stored patterns. Convergence times and the corresponding dispersions have also been studied numerically a.s para.meters to measure the efficiency of this model. These quantities are related to the number of spurious sLates for ea.ch configuration of stored pa.tterns. Both the average convergence time and ...
A model for neural networks is studied through a multineuron interaction energy function E given by the product of the squared distances in phase space between the state S of the net and the P stored patterns. Convergence times and the corresponding dispersions have also been studied numerically a.s para.meters to measure the efficiency of this model. These quantities are related to the number of spurious sLates for ea.ch configuration of stored pa.tterns. Both the average convergence time and the dispersions are small in the retrieval phase, indicating the homogeneity and smoothness of the phase space. ...
Institution
Universidade Federal do Rio Grande do Sul. Instituto de Física. Curso de Pós-Graduação em Física.
Collections
-
Exact and Earth Sciences (5188)Physics (842)
This item is licensed under a Creative Commons License
