Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

A Missing Link in CyberneticsBackpropagation

A Missing Link in Cybernetics: Backpropagation [Following the initiative of McCulloch and Pitts (1943), there has been much speculation about the achievement of artificial intelligence using networks of model neurons. The advent of the “perceptron” principle (Rosenblatt 1961; Nilsson 1965) crystallised something definite and functional out of a mass of diffuse speculation, but it is not difficult to show that the “simple perceptron” has limited capability (Minsky and Papert 1969). This can be attributed to the fact that all of the changes in weights constituting its learning are restricted to a single functional layer. The simple training algorithm is possible because all the places where changes occur are in this one layer and contribute directly to the output of the device, but the range of tasks that can be learned is drastically limited.] http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png

A Missing Link in CyberneticsBackpropagation

Part of the IFSR International Series on Systems Science and Engineering Book Series (volume 26)
Andrew, Alex M.
Springer Journals — Jan 27, 2009
Download PDF
19 pages

Loading next page...
 
/lp/springer-journals/a-missing-link-in-cybernetics-backpropagation-pwcKbbJTAW
Publisher
Springer New York
Copyright
© Springer-Verlag New York 2009
ISBN
978-0-387-75163-4
Pages
85 –104
DOI
10.1007/978-0-387-75164-1_5
Publisher site
See Chapter on Publisher Site

Abstract

[Following the initiative of McCulloch and Pitts (1943), there has been much speculation about the achievement of artificial intelligence using networks of model neurons. The advent of the “perceptron” principle (Rosenblatt 1961; Nilsson 1965) crystallised something definite and functional out of a mass of diffuse speculation, but it is not difficult to show that the “simple perceptron” has limited capability (Minsky and Papert 1969). This can be attributed to the fact that all of the changes in weights constituting its learning are restricted to a single functional layer. The simple training algorithm is possible because all the places where changes occur are in this one layer and contribute directly to the output of the device, but the range of tasks that can be learned is drastically limited.]

Published: Jan 27, 2009

Keywords: Hind Limb; Significance Feedback; Hide Unit; Cochlear Nucleus; Rubber Sheet

There are no references for this article.