Title :
A suggestion for solving negative transfer problem in VLSI backpropagation neural networks
Author :
Lo, Jien-Chung ; Kolluri, Srinivas
Author_Institution :
Dept. of Electr. Eng., Rhode Island Univ., Kingston, RI, USA
fDate :
27 Jun-2 Jul 1994
Abstract :
To overcome negative transfer problem in backpropagation (BP) neural networks, a modified BP algorithm has been proposed. However, for VLSI BP neural networks, this means an extensive design efforts. We analyze in this paper the nature of negative transfer problem and identify a simple technique to solve the negative transfer problem in VLSI BP neural networks
Keywords :
VLSI; backpropagation; neural chips; VLSI backpropagation neural networks; fault tolerance; learning characteristics; negative transfer problem; synaptic weights; Algorithm design and analysis; Backpropagation algorithms; Gaussian distribution; Histograms; Intelligent networks; Network-on-a-chip; Neural networks; Neurons; Performance analysis; Very large scale integration;
Conference_Titel :
Neural Networks, 1994. IEEE World Congress on Computational Intelligence., 1994 IEEE International Conference on
Conference_Location :
Orlando, FL
Print_ISBN :
0-7803-1901-X
DOI :
10.1109/ICNN.1994.374548
