Title :
Implementing backpropagation with analog hardware
Author_Institution :
Dept. of Electr. Eng., Eindhoven Univ. of Technol., Netherlands
fDate :
27 Jun-2 Jul 1994
Abstract :
The main problem when implementing backpropagation with analog hardware is the offset present in multipliers in the backward path of multilayer neural networks. Here this problem is further investigated and a solution is presented at the cost of the speed of the backpropagation algorithm
Keywords :
analogue integrated circuits; analogue multipliers; backpropagation; multilayer perceptrons; neural chips; analog hardware; backpropagation; backward path; multilayer neural networks; multipliers; Adders; Analog circuits; Backpropagation algorithms; Costs; Error correction; Multi-layer neural network; Neural network hardware; Neural networks; Neurons; Pulse width modulation;
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.1994.374522
