ABSTRACT
The chapter presents various types of supervised, unsupervised and reinforcement learning models built with artificial neural nets. Among the supervised models special emphasis has been given to Widrow-Hoff’s multi-layered ADALINEs and the back-propagation algorithm. The principles of unsupervised learning have been demonstrated through Hopfield nets, and the adaptive resonance theory (ART) network models. The reinforcement learning is illustrated with Kohonen’s self-organizing feature map. The concepts of fuzzy neural nets will also be introduced in this chapter to demonstrate its application in pattern recognition problems.
