Title :
A flexible architecture for neural networks
Author :
Ouali, J. ; Saucier, G.
Author_Institution :
Inst. Nat. Polytech. de Grenoble, CSI, France
Abstract :
A distributed, synchronous architecture for neural networks is presented. The basic processor is associated to one neuron and is able to autonomously perform all the steps of the learning and relaxation phases. Data circulation is implemented by shifting techniques. Customization of the network is done by fixing identification data in dedicated memory elements
Keywords :
neural nets; parallel architectures; data circulation; dedicated memory elements; distributed architecture; identification data; learning; network customization; neural networks; neuron; relaxation; shifting techniques; synchronous architecture; Computational modeling; Computer architecture; Computer simulation; Concurrent computing; Encoding; Information processing; Neural networks; Neurons; Silicon; Switches;
Conference_Titel :
Computer Design: VLSI in Computers and Processors, 1989. ICCD '89. Proceedings., 1989 IEEE International Conference on
Conference_Location :
Cambridge, MA
Print_ISBN :
0-8186-1971-6
DOI :
10.1109/ICCD.1989.63413
