Title :
A localized learning rule for analog VLSI implementation of neural networks
Author :
Wasaki, Hiroyuki ; Horio, Yoshihiko ; Nakamura, Shogo
Author_Institution :
Dept. of Electron. Eng., Tokyo Denki Univ., Japan
Abstract :
A modified Hebbian type learning rule for self-organization which uses only the local information is proposed. As the result of computer simulations of self-organizing networks, the validity of the rule was confirmed and learning speed was improved. Furthermore, circuit examples for implementing the learning rule are proposed
Keywords :
VLSI; analogue circuits; learning systems; neural nets; analog VLSI implementation; computer simulations; learning speed; local information; localized learning rule; modified Hebbian type learning rule; neural networks; self-organization; self-organizing networks; Circuit simulation; Computational modeling; Computer simulation; Distributed processing; Hardware; Integrated circuit interconnections; Neural networks; Neurons; Very large scale integration; Wires;
Conference_Titel :
Circuits and Systems, 1990., Proceedings of the 33rd Midwest Symposium on
Conference_Location :
Calgary, Alta.
Print_ISBN :
0-7803-0081-5
DOI :
10.1109/MWSCAS.1990.140641
