Title :
A new realization of the conscience mechanism for self-organizing neural networks implemented in CMOS technology
Author :
TalasÌka, Tomasz ; Rydlewski, Andrzej ; DÅ‚ugosz, RafaÅ‚
Author_Institution :
Fac. of Telecommun. & Electr. Eng., Univ. of Technol. & Life Sci., Bydgoszcz, Poland
Abstract :
The paper presents a simple current-mode conscience mechanism used in hardware implemented Winner Takes All neural networks to eliminate the problem of the, so called, dead neurons. These neurons reduce the number of data classes that can be distinguished by the NN, i.e. they decrease the efficiency of the NN. The circuit has been realized in the TSMC CMOS 0.18 μm process. At data rate of about 10 MSamples/s it dissipates the power of 15 μW.
Keywords :
CMOS integrated circuits; application specific integrated circuits; self-organising feature maps; CMOS technology; current mode conscience mechanism; dead neurons; power 15 muW; self organizing neural networks; size 0.18 mum; Artificial neural networks; Biological neural networks; CMOS integrated circuits; CMOS technology; Hardware; Neurons; Transistors;
Conference_Titel :
Microelectronics Proceedings - MIEL 2014, 2014 29th International Conference on
Conference_Location :
Belgrade
Print_ISBN :
978-1-4799-5295-3
DOI :
10.1109/MIEL.2014.6842190
