Selfrefreshing analogue memory cell for variable synaptic weights
作者： R. Castello , D.D. Caviglia , M. Franciotta , F. Montecchi
DOI: 10.1049/EL:19911161
关键词: Memory cell 、 Capacitor 、 Computer science 、 Signal 、 Very-large-scale integration 、 Topology 、 Electronic engineering 、 Variable (computer science) 、 Artificial neural network 、 CMOS
摘要: An analogue memory cell based on a local selfrefreshing technique is presented. The circuit suitable for long-term storage of synaptic weights in VLSI implementations neural networks. refreshes the voltage capacitor multilevel fashion by automatic comparison and confirmation with range discrete levels locally reproduced from an external reference signal.
core.ac.uk
sci-hub.se 参考文章(3)
Holler, Tam, Castro, Benson, An electrically trainable artificial neural network (ETANN) with 10240 'floating gate' synapses international joint conference on neural network. pp. 191- 196 ,(1989) , 10.1109/IJCNN.1989.118698
E.A. Vittoz, Analog VLSI implementation of neural networks international symposium on circuits and systems. pp. 2524- 2527 ,(1990) , 10.1109/ISCAS.1990.112524
T. Morishita, Y. Tamura, T. Otsuki, A BiCMOS analog neural network with dynamically updated weights international solid-state circuits conference. pp. 142- 143 ,(1990) , 10.1109/ISSCC.1990.110167