An Introduction to Neural Networks

An Introduction to Neural Networks falls into a new ecological niche for texts. Basedon notes that have been class-tested for more than a decade, it is aimed at cognitive science andneuroscience students who need to understand brain function in terms of computational modeling, andat engineers who want to go beyond formal algorithms to applications and computing strategies. It isthe only current text to approach networks from a broad neuroscience and cognitive scienceperspective, with an emphasis on the biology and psychology behind the assumptions of the models, aswell as on what the models might be used for. It describes the mathematical and computational toolsneeded and provides an account of the author's own ideas.Students learn how to teach arithmetic to aneural network and get a short course on linear associative memory and adaptive maps. They areintroduced to the author's brain-state-in-a-box (BSB) model and are provided with some of theneurobiological background necessary for a firm grasp of the general subject.The field now known asneural networks has split in recent years into two major groups, mirrored in the texts that arecurrently the engineers who are primarily interested in practical applications of the newadaptive, parallel computing technology, and the cognitive scientists and neuroscientists who areinterested in scientific applications. As the gap between these two groups widens, Anderson notesthat the academics have tended to drift off into irrelevant, often excessively abstract researchwhile the engineers have lost contact with the source of ideas in the field. Neuroscience, he pointsout, provides a rich and valuable source of ideas about data representation and setting up the datarepresentation is the major part of neural network programming. Both cognitive science andneuroscience give insights into how this can be done cognitive science suggests what tocompute and neuroscience suggests how to compute it.