期刊
IEEE PHOTONICS JOURNAL
卷 14, 期 6, 页码 -出版社
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/JPHOT.2022.3219246
关键词
Optical neural network; microring resonator; all-optical nonlinear activation functions; thermo-optical effect
资金
- National Key Research and Development Program of China [2019YFB2203303]
- National Natural Science Foundation of China [62071042]
- National Natural Science Foundation of China (NSFC) [62105028]
- China Postdoctoral Science Foundation [2021M690390]
This paper proposes and experimentally demonstrates a photonic method to implement reconfigurable and low-threshold all-optical nonlinear activation functions based on a compact and high-Q add-drop microring resonator on silicon. The experiment shows that various nonlinear activation functions can be realized with low threshold and high accuracy, making it suitable for large-scale integrated optical neural networks (ONNs).
The realization of optical nonlinear activation functions (NAFs) is essential for integrated optical neural networks (ONNs). Here, we propose and experimentally demonstrate a photonic method to implement reconfigurable and low-threshold all-optical NAFs based on a compact and high-Q add-drop microring resonator (MRR) on silicon. In the experiment, four different NAFs including softplus, radial basis, clamped ReLU, and sigmoid functions are realized by exploiting the thermo-optical (TO) effect of the MRR. The threshold to implement NAFs is as low as 0.08 mW. As a demonstration, a handwritten digit classification benchmark task is simulated based on a convolutional neural network (CNN) using the obtained activation functions, where an accuracy of 98% is realized. Thanks to the unique advantages of ultra-compact footprint and ultralow threshold, the proposed nonlinear unit is promising to be widely used in large-scale integrated ONNs.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据