High-Speed Optical Convolutional Neural Network Accelerator with 100 Gbaud EO-polymer/Si Hybrid Optical Modulator
Akito Shinya, Koji Kida, Hiromu Sato, Guo-Wei Lu, Shiyoshi Yokoyama, Junichi Fujikata · 2023
Convolutional neural networks are a powerful category of artificial neural networks that can extract hierarchical features of biological data to significantly reduce parametric complexity and increase prediction accuracy. In this paper, we propose high-speed optical convolutional neural network accelerator with 100Gbaud EO-polymer/Si hybrid optical modulator and demonstrated proof of concept both in simulation and experiment.