Application of Hopfield-Like Neural Networks to Linear Factorization

0404680 - UIVT-O 20020157 RIV DE eng C - Conference Paper (international conference)
Húsek, Dušan - Frolov, A. A. - Řezanková, H. - Snášel, Václav
Application of Hopfield-Like Neural Networks to Linear Factorization.
COMPSTAT 2002. Proceedings in Computational Statistics. Heidelberg: PhysicaVerlag, 2002 - (Härdle, W.; Rönz, B.), s. 177-182. ISBN 3-7908-1517-9.
[COMPSTAT 2002. Berlin (DE), 24.08.2002-28.08.2002]
R&D Projects: GA ČR GA201/01/1192; GA ČR GA201/00/1031
Institutional research plan: AV0Z1030915
Keywords : binary factorization * Hopfield network * sparse encoding
Subject RIV: BB - Applied Statistics, Operational Research

The problem of binary factorization of complex patterns in reccurent Hopfield-like neural network was studied by means of computer simulation. The network ability to perform a factorization was analyzed depending on the number and sparsness of factors mixed in present patterns. Binary factorization in sparsely encoded Hopfield-like network is treated as efficient statistical method and as a functional model of hippocampal CA3 field.
Permanent Link: http://hdl.handle.net/11104/0124919