Why the fuss about neural networks?

dc.contributor.author	Barth, G
dc.date.accessioned	2018-05-19T00:35:52Z
dc.date.available	2018-05-19T00:35:52Z
dc.date.issued	1991
dc.identifier.citation	Barth G (1991) Why the fuss about neural networks? South African Computer Journal, Number 5, 1991	en
dc.identifier.issn	2313-7835
dc.identifier.uri	http://hdl.handle.net/10500/23944
dc.description.abstract	Currently, there is a certain hype about issues related to neural networks. This is somewhat surprising since the origins of neural nets date back some 40 years. At that time, several researchers experimented with distributed systems composed of so-called perceptrons. Problems tackled therewith included computer vision, weather prediction, speech recognition, cardiographic diagnosis and image classification. Chapter 2 of reference [1] contains more details. This short contribution is intended to summarize in a very succinct way the main characteristics of neural nets as currently used. In particular, the two dominating training paradigms are briefly described. An attempt is made to sketch some guidelines as to what kinds of problems are good candidates for being tackled by neural networks.	en
dc.language		en
dc.language.iso	en	en
dc.publisher	South African Institute of Computer Scientists and Information Technologists	en
dc.title	Why the fuss about neural networks?	en
dc.type	Article	en