Title :
Integration of magnified gradient function and weight evolution with deterministic perturbation into back-propagation
Author :
Ng, Sin-Chun ; Cheung, Chi-Chung ; Leung, Shu-hung
Author_Institution :
Sch. of Sci. & Technol., Open Univ. of Hong Kong, Kowloon, China
fDate :
3/6/2003 12:00:00 AM
Abstract :
An integrated approach of magnified gradient function and weight evolution with deterministic perturbation to improve the performance of back-propagation learning is proposed. Simulation results show that, in terms of the convergence rate and the percentage of global convergence, the integrated approach always outperforms the other traditional methods.
Keywords :
backpropagation; convergence; back-propagation algorithm; backpropagation learning; convergence rate; deterministic perturbation; global convergence; magnified gradient function; performance improvement; weight evolution;
Journal_Title :
Electronics Letters
DOI :
10.1049/el:20030277
