On the ability of the optimal perceptron to generalise

, , and

Published under licence by IOP Publishing Ltd
, , Citation M Opper et al 1990 J. Phys. A: Math. Gen. 23 L581 DOI 10.1088/0305-4470/23/11/012

Download Article PDF

Article metrics

392 Total downloads

Share this article

Author affiliations

1 Institut fur Theor. Phys., Justus-Liebig Univ., Giessen, West Germany

Buy this article in print

0305-4470/23/11/L581

Abstract

A linearly separable Boolean function is derived from a set of examples by a perceptron with optimal stability. The probability to reconstruct a pattern which is not learnt is calculated analytically using the replica method.

Export citation and abstract BibTeX RIS

Previous article in issue
Next article in issue
10.1088/0305-4470/23/11/012