autonomous vehicle control). Finally, extensions of backpropogation for temporal learning (time delay neural networks) and recurrent neural networks are discussed. These topics are important in the neural network literature and Hassoun is successful in explaining the basis of these algorithms. Chapter 6 introduces several additional adaptive multilayer networks and their associated training procedures, including Radial Basis Functions (RBF) networks and the Cerebellar Model Articulation Model (CMAC). RBF networks are motivated by biological nervous systems, and their advantages and disadvantages with respect to backpropogation are discussed. H that accuracy can be improved through two major variations of backpropogation. The next network, CMAC, is another example of a localized receptive field function for each. Several methods to improve capacity and error correction in these networks are discussed, and a variety of other DAM are introduced: brain-state-in-a-box model, nonmonotonic activations model, hysteretic activations model and heteroassociative model. Chapter 8 discusses global search methods applied to multilayer networks: stochastic simulated annealing, mean field annealing, and genetic algorithms. Finally, an improved hybrid genetic algorithm/gradient search method for feedforward neural net training is presented. Hassoun has written a good text- rice ensive of adaptive multilayer networks addressed, ART1 and the autoassociative clustering network, have
/lp/association-for-computing-machinery/book-review-analysis-and-applications-of-artificial-neural-networks-by-MgfldAtPgN