Getting started -- Perceptron learning -- basics -- Choice of learning rules -- Augmented statistical mechanics formulation -- Noisy teachers -- Storage problem -- Discontinuous learning -- Unsupervised learning -- On-line learning -- Making contact with statistics -- Bird's eye view: multifractals -- Multilayer networks -- On-line learning in multilayer networks -- What else?
Summary
Artificial neural networks provide a simple framework for describing learning from examples. This coherent account of important concepts and techniques of statistical mechanics and their application to learning theory comes with background material in mathematics and physics, plus many examples and exercises, making it ideal for courses, self-teaching, or reference
Bibliography
Includes bibliographical references (pages 313-325) and index