[23a-M206-1~10] FS.1 Focused Session "AI Electronics"

[23a-M206-2] Magneto-Optical Diffractive Deep Neural Network

〇(M2)Takumi Fujita¹, Hotaka Sakaguchi¹, Jian Zhang¹, Hirofumi Nonaka², Satoshi Sumi³, Hiroyuki Awano³, Takayuki Ishibashi¹ (1.Nagaoka Univ. Tech., 2.Aichi Inst. Tech., 3.Toyota Tech. Inst.)

Keywords:neural network, optical computing, magneto-optics

In recent years, the rapid development of Deep Neural Network (" DNN ") has led to its application in a variety of fields. However, the processing speed and power consumption increase significantly in proportion to the complexity of their models. To solve these problems, physically implemented DNN that operate at high speed and low power consumption are required. We propose an optical diffractive DNN based on the magneto-optic effect, which operates in visible light and is highly compatible with current imaging devices.