BATCH NORMALIZATION APPARATUS BASED ON RESISTIVE MEMORY FOR BINARY NEURAL NETWORK

摘要

The artificial neural network circuit according to the present embodiment includes a weight memory device that stores weight information, a unit weight memory cell including a weight pass transistor, a threshold memory device that is programmed to store a threshold, and a threshold pass. A unit threshold memory cell including a transistor, a weight-threshold column in which a plurality of unit weight memory cells and a plurality of unit threshold memory cells are connected, and an output signal of the weight-threshold column are provided as an input. , And a sense amplifier in which the reference voltage is provided to the other input.