Neural Nets

doi:10.1117/3.2569178.ch5

Neural Nets

Book: An Introduction to Optics in Computers

Author(s): Henri H. Arsenault, Yunlong Sheng

Published: 1992

https://doi.org/10.1117/3.2569178.ch5

Author Affiliations +

Abstract

Artificial neural nets are networks containing a large number of relatively simple processors, roughly modeled on the brain. In a neural net, the complexity resides in the interconnections between neurons, which are usually connected in layers. Neural net science is not a new field, but it stagnated after the end of the sixties following the publication of a book by Minsky and Papert showing that perceptrons (two-layer neural nets) could only carry out a limited number of operations. The field revived in the early eighties after Hopfield showed that some of the problems of perceptrons could be alleviated by using more than two layers.