Turchetti (Universita Politecnica delle Marche) investigates the properties of neural networks as sources of random functions, and whether approximating properties similar to those valid for deterministic functions hold for random functions. After introducing the mathematical model of artificial neural networks, he describes some fundamental results of the stochastic process theory, examines the space of functions generated by stochastic neural networks, and defines architectures for the implementation of stochastic networks. The final chapter considers the process of modeling and memorizing physical events as stochastic processes, and approaches the problem of neural computing from a different angle than digital computers. Annotation ©2005 Book News, Inc., Portland, OR (booknews.com)
