DocumentCode :
3429613
Title :
Convergence mechanism of learning on backpropagation and high speed learning
Author :
Sakai, A. ; Iijima, N. ; Yoshida, Y. ; Mitsui, H. ; Sone, M.
Author_Institution :
Musashi Inst. of Technol., Tokyo, Japan
fYear :
1992
fDate :
16-20 Nov 1992
Firstpage :
523
Abstract :
The input signals of units on neural network (NN) are converted to output signals by an I/O function such as the sigmoid functions. The input signals at each layer unit is changed by the learning. Its change is dependent on the slope of the I/O function. In the sensitive region (active region) given by the I/O function, input signals move effectively and the movement increase with the increase of the slope. At the saturation region, however, it is very small. The limited linear function is used instead of the sigmoid function, because the active region is defined clearly and an examination of the movement is made easily. The convergence mechanism of learning by back propagation is examined by the movement per training cycle. The possibility of the high speed learning is considered on the basis of the results
Keywords :
backpropagation; convergence; neural nets; I/O function; active region; backpropagation; convergence mechanism; high speed learning; input signals; linear function; neural network; output signals; saturation region; sensitive region; Backpropagation; Convergence; Joining processes; Neural networks;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Singapore ICCS/ISITA '92. 'Communications on the Move'
Print_ISBN :
0-7803-0803-4
Type :
conf
DOI :
10.1109/ICCS.1992.254896
Filename :
254896
Link To Document :
بازگشت