All-optical ultrafast ReLU function for energy-efficient nanophotonic deep learning

被引：48

作者：

Li, Gordon H. Y. ^{[1
]}

Sekine, Ryoto ^{[2
]}

Nehra, Rajveer ^{[2
]}

Gray, Robert M. ^{[2
]}

Ledezma, Luis ^{[1
,3
]}

Guo, Qiushi ^{[2
]}

Marandi, Alireza ^{[1
,2
]}

机构：

[1] CALTECH, Dept Appl Phys, Pasadena, CA 91125 USA

[2] CALTECH, Dept Elect Engn, Pasadena, CA 91125 USA

[3] CALTECH, Jet Prop Lab, Pasadena, CA 91125 USA

来源：

NANOPHOTONICS | 2023年 / 12卷 / 05期

基金：

美国国家科学基金会;

关键词：

deep learning; optical computing; optical neural networks; thin-film lithium niobate; ACTIVATION FUNCTION; NEURAL-NETWORKS; PERFORMANCE; CONVERSION;

D O I：

10.1515/nanoph-2022-0137

中图分类号：

TB3 [工程材料学];

学科分类号：

0805 ; 080502 ;

摘要：

In recent years, the computational demands of deep learning applications have necessitated the introduction of energy-efficient hardware accelerators. Optical neural networks are a promising option; however, thus far they have been largely limited by the lack of energy-efficient nonlinear optical functions. Here, we experimentally demonstrate an all-optical Rectified Linear Unit (ReLU), which is the most widely used nonlinear activation function for deep learning, using a periodically-poled thin-film lithium niobate nanophotonic waveguide and achieve ultra-low energies in the regime of femtojoules per activation with near-instantaneous operation. Our results provide a clear and practical path towards truly all-optical, energy-efficient nanophotonic deep learning.

引用

页码：847 / 855

页数：9

共 52 条

[1] ITO-based electro-absorption modulator for photonic neural activation function [J].