Optical Implementations of Neural Computing

Abstract

Achieving performance comparable to human beings in speech recognition, visual perception, motor control and knowledge acquisition, representation, and processing is one of the most difficult and exciting challenges facing the information processing research community. Recently neural net models of computation have been investigated as a novel approach for solving these problems. These proposed models are only loosely based on the known and postulated characteristics of biological systems and no claim is usually made for these models to be biologically accurate.