Working principle of the sensing-memory-computing device. (IMAGE)
Caption
a, Single-layer perceptron (SLP) neural network architecture schematics and the accuracy of the SLP classifier during training with floating-point weight and discrepant 7-level photoresponse state. b, Illustration of an in-memory IMP operation based on a set of photomemristors triggered by light stimuli. c, Diagram of the optical and electrical pulses applied for the IMP operation. The blue and red curves show the optical and electrical signal of devices p and q before and during the IMP operation, and the orange curves show the change in photoresponse states after IMP operations, reproducing the IMP truth table.
Credit
by Xiao Fu, Tangxin Li, Bin Caid, Jinshui Miao, Gennady N. Panin, Xinyu Ma, Jinjin Wang, Xiaoyong Jiang, Qing Lia, Yi Dong, Chunhui Hao, Juyi Sun, Hangyu Xu, Qixiao Zhao, Mengjia Xia, Bo Song, Fansheng Chen, Xiaoshuang Chen, Wei Lu, Weida Hu
Usage Restrictions
Credit must be given to the creator.
License
CC BY