POWER-LAW SLOWDOWN OF THE NEURAL LEARNING

Authors
CATEAU H NAKAJIMA T NUNOKAWA H FUCHIKAMI N
Citation
H. Cateau et al., POWER-LAW SLOWDOWN OF THE NEURAL LEARNING, IEICE transactions on fundamentals of electronics, communications and computer science, E77A(12), 1994, pp. 2109-2111
Citations number
NO
Categorie Soggetti
Engineering, Eletrical & Electronic","Computer Science Hardware & Architecture","Computer Science Information Systems
Journal title
IEICE transactions on fundamentals of electronics, communications and computer science → ACNP
ISSN journal
09168508
Volume
E77A
Issue
12
Year of publication
1994
Pages
2109 - 2111
Database
ISI
SICI code
0916-8508(1994)E77A:12<2109:PSOTNL>2.0.ZU;2-3
Abstract
We numerically show that the learning time t of the back propagation m odel with the encoder topology obeys a power law described as t propor tional to M(D) (D: constant, 1 < D less than or similar to 2), with M being the number of input patterns.