VLSI realization of artificial neural Networks using digital and mixed Signal hardware

Abstract

A Biological Neural Network BNN forms the Central Nervous newlineSystem which has highly interconnected neurons to co ordinate all the newlinefunctions like reading and thinking Artificial Neurons are simple newlineabstractions of biological neurons programmed in software or modeled in newlinehardware Networks of artificial neurons known as Artificial Neural Networks newline ANN have a fraction of power of biological neural structures and they can newlinebe trained to solve many complex problems An ANN has the ability to learn newlineand it had been realized in software But learning is a recursive process newlineinvolving multiple iterations which takes a long time when implemented in newlinesoftware However availability of powerful resources at affordable cost has newlineopened up the interesting possibility of realizing ANN in hardware The focus newlineof the present work is on Very Large Scale Integration VLSI hardware newlinerealization of ANN so as to achieve flexibility portability and adaptability newline newline