Title :
Hardware realisation of artificial neural network with application to information coding
Author :
Young, D. ; Lee, A.W.H. ; Cheng, LM
Author_Institution :
Dept. of Electron. Eng., City Polytech. of Hong Kong, Kowloon, Hong Kong
fDate :
27 Jun- 2 Jul 1994
Abstract :
A general purpose flexible and low cost ANN system has been designed and built. The hardware is based on four MD1220 neural bit slice processors interfaced to a PC/AT host computer. The system can be configured to run concurrently as one partition or up to 4 independent partitions, with dedicated firmware control and weight placement in the Local Weight Ram of each MD1220. The hardware is capable of storing up to 65536 16 bit weights with its internal 256 K RAM and 1024 neurons with its global RAM using time division multiplexing technique. Back propagation learning paradigm is used for evaluating the system and application of (7,4) Hamming coding is demonstrated
Keywords :
Hamming codes; backpropagation; bit-slice computers; microcomputer applications; neural chips; time division multiplexing; (7,4) Hamming coding; 16 bit; 256 KB; Local Weight Ram; MD1220 neural bit slice processors; PC/AT host computer; TDM; artificial neural network; backpropagation learning paradigm; dedicated firmware control; hardware realisation; information coding; time division multiplexing technique; weight placement; Application software; Artificial neural networks; Computer interfaces; Control systems; Costs; Microprogramming; Neural network hardware; Neurons; Read-write memory; Weight control;
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.1994.374849
