Implementation of an analog self-learning neural network
Chun Lu, Bingxue Shi, Chen Lü · 2002
An analog self-learning neural network is proposed. A prototype chip has been fabricated using a 1.2-um CMOS, double-polysilicon, double-metal technology. Because of its fully analog and fully parallel structure, the proposed LSI can do continuous time calculation. The result of the XOR experiment shows that the circuit achieves the self-learning.