Department of Electronics Engineering, Xiamen University, P.O. Box 802, Xiamen, Fujian 361005, China
Abstract:
A new method is proposed to represent the bipolar states of a neural network. An optoelectronic system based on an inner-product multiplier is suggested to implement the bipolar associative memory which has a larger storage capacity and a higher error-correction capability than those of unipolar one. The experimental results are also given and discussed.