Title :
Analog neural network circuits for ASIC fabrication
Author :
Dundar, G. ; Rose, K.
Author_Institution :
Center for Integrated Electron., Rensselaer Polytech. Inst., Troy, NY, USA
Abstract :
The design of high speed analog neural network (ANN) circuits suitable for ASIC fabrication is presented. The dependence of ANN performance in pattern recognition on device, circuit, and network characteristics is demonstrated. Alternative electrically alterable read only memory (EAROM) technologies for implementing analog neural networks are compared
Keywords :
CMOS integrated circuits; EPROM; analogue processing circuits; application specific integrated circuits; linear integrated circuits; neural chips; ASIC fabrication; EAROM; analog neural network; electrically alterable read only memory; high speed; neuron circuit; pattern recognition; synapse circuit; Application specific integrated circuits; Artificial neural networks; Biological neural networks; EPROM; Fabrication; Humans; Integrated circuit interconnections; MOSFETs; Neural networks; Voltage;
Conference_Titel :
ASIC Conference and Exhibit, 1992., Proceedings of Fifth Annual IEEE International
Conference_Location :
Rochester, NY
Print_ISBN :
0-7803-0768-2
DOI :
10.1109/ASIC.1992.270256
