Title :
Adaptive training and pruning in feedforward networks
Author :
Chang, Sheng-Jiang ; Sum, John ; Wong, Kwok-Wo ; Leung, Chi-sing
Author_Institution :
Inst. of Modern Opt., Nankai Univ., Tianjin, China
fDate :
1/18/2001 12:00:00 AM
Abstract :
A local extended Kalman filter training and pruning approach is proposed to train feedforward networks with the goal of reducing the computational complexity and storage requirement in large-scale practical problems. The performance of the proposed algorithm is demonstrated for the problem of handwritten digit recognition
Keywords :
Kalman filters; computational complexity; feedforward neural nets; handwritten character recognition; learning (artificial intelligence); adaptive pruning; adaptive training; computational complexity reduction; feedforward networks; handwritten digit recognition; local extended Kalman filter approach; storage requirement reduction;
Journal_Title :
Electronics Letters
DOI :
10.1049/el:20010074
