講演情報
[9p-N106-8]Monte Carlo Optimization for Online Learning in Magneto-Optical Diffractive Deep Neural Network
〇FatimaZahra Chafi1, Tomonao Matsuya1, Hotaka Sakaguchi1, Hirofumi Nonaka2, Hiroyuki Awano3, Takayuki Ishibashi1 (1.Nagaoka Univ. Tech., 2.Aichi Inst. Tech., 3.Toyota Tech. Inst.)
キーワード:
Monte Carlo Optimization、MO-D2NN、Online Learning
Recent advances in physical neural networks, particularly magneto-optical diffractive deep neural network (MO-D2NN), have demonstrated remarkable potential for performing complex inference tasks at high-speed, and with minimal energy consumption. However, in order to reduce the power consumption during the training, a learning method with low computational cost is required. In this work, we propose an alternative approach based on Monte Carlo learning algorithm, tailored for online learning in MO-D2NN.