Information Storage in Hopfield Model with Reduced Complexity


We developed a dynamic neural network with block triangular interconnection weight matrix for associative memory designs. This model has equivalent storage capacity as a fully connected network. In other words, we show the following: For a Hopfield neural network, = x + Wy + I, y = f (x ) , vector pattern set E q = {y~, y2 , . . . , yq}, yi E •P, 1 ~< i… (More)
DOI: 10.1016/S0020-0255(98)10012-9


4 Figures and Tables

Slides referencing similar topics