Why VLSI Implementations of Associative VLCNs Require Mult iplexingt

In recent years there has been a great deal of interest developed ’ Connectionist/Neural Networks (CNNb; These are networks of simple processors where the network state is represented by the outputs of the processors and the strengths of the connections between them. They are characterized by asynchrony, a high degree of connectivity, and good fault tolerance. CNNs are seen as providing solutions to problems that are difficult or impossible to solve with the traditional approach of symbolic processing on von Neumann computers. Pattern recognition, constraint satisfaction, and data completion are some examples of these types of problems.