DocumentCode :
2066764
Title :
An analogue neural network using MCM technology
Author :
Jabri, Marwan ; Leong, Philip ; Burr, Jim ; Flower, Barry ; Lai, Kam K. ; Pickard, Stephen ; Tinker, Edward ; Coggins, Richard
Author_Institution :
SEDAL, Sydney Univ., NSW, Australia
fYear :
1993
fDate :
24-26 Nov 1993
Firstpage :
122
Lastpage :
125
Abstract :
Describes a large analogue neural network implemented using multichip module (MCM) technology to provide interchip connections. An array of eight synapse chips, each containing 2500 synapses, is connected to a neuron chip on an MCM to implement a (150, 100, 50) network containing more than one million transistors. MCM technology enables a higher synapse density to be achieved and benefits which include smaller area and reduced parasitics can be enjoyed. The MCM is under construction, and its completion will provide a large analogue neural network in a single package which can be used for many applications which require such networks
Keywords :
analogue computer circuits; linear integrated circuits; multichip modules; neural chips; analogue neural network; area; interchip connections; multichip module technology; neuron chip; parasitics; synapse chips; synapse density; Artificial neural networks; Integrated circuit interconnections; Multi-layer neural network; Multichip modules; Multilayer perceptrons; Neural networks; Neurons; Very large scale integration; Voltage; Wire;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Artificial Neural Networks and Expert Systems, 1993. Proceedings., First New Zealand International Two-Stream Conference on
Conference_Location :
Dunedin
Print_ISBN :
0-8186-4260-2
Type :
conf
DOI :
10.1109/ANNES.1993.323066
Filename :
323066
Link To Document :
بازگشت