On-line learning of non-monotonic rules by simple perceptron
Abstract
We study the generalization ability of a simple perceptron which learns unlearnable rules. The rules are presented by a teacher perceptron with a non-monotonic transfer function. The student is trained in the on-line mode. The asymptotic behaviour of the generalization error is estimated under various conditions. Several learning strategies are proposed and improved to obtain the theoretical lower bound of the generalization error.
- Publication:
-
Journal of Physics A Mathematical General
- Pub Date:
- June 1997
- DOI:
- 10.1088/0305-4470/30/11/012
- arXiv:
- arXiv:cond-mat/9703020
- Bibcode:
- 1997JPhA...30.3795I
- Keywords:
-
- Condensed Matter;
- Nonlinear Sciences - Adaptation and Self-Organizing Systems
- E-Print:
- LaTeX 20 pages using IOP LaTeX preprint style file, 14 figures