𝔖 Bobbio Scriptorium
✦   LIBER   ✦

A modified back-propagation method to avoid false local minima

✍ Scribed by Yutaka Fukuoka; Hideo Matsuki; Haruyuki Minamitani; Akimasa Ishida


Publisher
Elsevier Science
Year
1998
Tongue
English
Weight
218 KB
Volume
11
Category
Article
ISSN
0893-6080

No coin nor oath required. For personal study only.

✦ Synopsis


The back-propagation method encounters two problems in practice, i.e., slow learning progress and convergence to a false local minimum. The present study addresses the latter problem and proposes a modified back-propagation method. The basic idea of the method is to keep the sigmoid derivative relatively large while some of the error signals are large. For this purpose, each connecting weight in a network is multiplied by a factor in the range of (0,1], at a constant interval during a learning process. Results of numerical experiments substantiate the validity of the method.