Title :
A probabilistic model of neural networks with static attractors
Author :
Yamanaka, Kazuo ; Agu, Masahiro
Author_Institution :
Dept. of Precision Eng., Ibaraki Univ., Hitachi, Japan
Abstract :
A probabilistic version of the binary Hopfield networks is proposed. Operation of the network is completely parallel, in the sense that evolution of each unit is governed only by its inherent probabilistic law. It is shown that the global state is attracted by one of the equilibria with probability one
Keywords :
neural nets; parallel processing; probability; binary Hopfield networks; global state; neural networks; probabilistic model; probability; static attractors; Neural networks; Precision engineering; Probability; State-space methods;
Journal_Title :
Systems, Man and Cybernetics, IEEE Transactions on
