[18p-A410-1~8] FS.1 Focused Session "AI Electronics"

[18p-A410-7] Development of fabrication process of magneto-optical diffractive deep neural networks

〇Hotaka Sakaguchi¹, Takumi Fujita¹, Jian Zhang¹, Satoshi Sumi², Hiroyuki Awano², Hirofumi Nonaka³, Takayuki Ishibashi¹ (1.Nagaoka Univ. of Tech., 2.Toyota Tech. Inst., 3.Aichi Inst. of Tech.)

Keywords:optical neural network, magneto-optical effect, deep neural network

Deep neural networks have been applied in various applications such as image recognition and speech recognition, but problems have occurred that increase processing speed and power consumption. Physical devices with high speed and low power consumption are expected to solve these problems. We have proposed a magneto-optical diffractive deep neural network (MO-D²NN), and reported that it is feasible with visible light operation and existing magneto-optical materials and magnetic domain formation technology. In this report, we report a prototype MO-D²NN with two hidden layers.