Premium
A neural network unification
Author(s) -
Aarts E. H. L.,
Ten Eikelder H. M. M.
Publication year - 2000
Publication title -
statistica neerlandica
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.52
H-Index - 39
eISSN - 1467-9574
pISSN - 0039-0402
DOI - 10.1111/1467-9574.00140
Subject(s) - boltzmann machine , hopfield network , boltzmann constant , restricted boltzmann machine , artificial neural network , relation (database) , unification , function (biology) , computer science , mathematics , energy (signal processing) , statistical physics , artificial intelligence , physics , data mining , quantum mechanics , statistics , evolutionary biology , biology , programming language
We study the relation between the asymptotic behaviour of synchronous Boltzmann machines and synchronous Hopfield networks. More specifically, we consider the relation between the pseudo consensus function that is used in analyzing Boltzmann machines and the energy function that is used in the study of Hopfield networks. We show that for small values of the control parameter, synchronous Boltzmann machines and synchronous Hopfield networks compute global respectively local maxima of the same function.