Title :
Initial weight selection methods for self-organizing training
Author_Institution :
Dept. of Autom. Eng., Beijing Inst. of Light Ind., China
Abstract :
Two initial weight selection methods for self organizing training of neural nets are introduced. These methods can improve the chances for successful training and speed up training processes remarkably. Two successful examples are given to show that these weight selection methods work well
Keywords :
learning (artificial intelligence); neural nets; self-adjusting systems; initial weight selection methods; neural nets; self organizing training; Aerospace industry; Automation; Belts; Industrial training; Light scattering; Neurons; Organizing; Vectors;
Conference_Titel :
Intelligent Processing Systems, 1997. ICIPS '97. 1997 IEEE International Conference on
Conference_Location :
Beijing
Print_ISBN :
0-7803-4253-4
DOI :
10.1109/ICIPS.1997.672810
