Space-efficient optical computing with an integrated chip diffractive neural network.

Autor: Zhu HH; Quantum Science and Engineering Centre (QSec), Nanyang Technological University, Singapore, 639798, Singapore., Zou J; Quantum Science and Engineering Centre (QSec), Nanyang Technological University, Singapore, 639798, Singapore., Zhang H; Quantum Science and Engineering Centre (QSec), Nanyang Technological University, Singapore, 639798, Singapore., Shi YZ; National Key Laboratory of Science and Technology on Micro/Nano Fabrication, Department of Micro/Nano Electronics, Shanghai Jiao Tong University, Shanghai, 200240, China., Luo SB; Quantum Science and Engineering Centre (QSec), Nanyang Technological University, Singapore, 639798, Singapore., Wang N; Institute of Microelectronics, A*STAR (Agency for Science, Technology and Research), Singapore, 138634, Singapore., Cai H; Institute of Microelectronics, A*STAR (Agency for Science, Technology and Research), Singapore, 138634, Singapore., Wan LX; Quantum Science and Engineering Centre (QSec), Nanyang Technological University, Singapore, 639798, Singapore., Wang B; Quantum Science and Engineering Centre (QSec), Nanyang Technological University, Singapore, 639798, Singapore., Jiang XD; Quantum Science and Engineering Centre (QSec), Nanyang Technological University, Singapore, 639798, Singapore. exdjiang@ntu.edu.sg., Thompson J; Centre for Quantum Technologies, National University of Singapore, Singapore, 117543, Singapore., Luo XS; Advanced Micro Foundry, 11 Science Park Road, 117685, Singapore, Singapore., Zhou XH; State Key Joint Laboratory of ESPC, Center for Sensor Technology of Environment and Health, School of Environment, Tsinghua University, Beijing, 100084, China. xhzhou@mail.tsinghua.edu.cn., Xiao LM; Shanghai Engineering Research Center of Ultra-Precision Optical Manufacturing, School of Information Science and Technology, Fudan University, Shanghai, 200433, China. liminxiao@fudan.edu.cn., Huang W; Suzhou Institute of Nano-Tech and Nano-Bionics (SINANO), Chinese Academy of Sciences (CAS), Suzhou, 215123, China., Patrick L; Advanced Micro Foundry, 11 Science Park Road, 117685, Singapore, Singapore., Gu M; Quantum Hub, School of Physical and Mathematical Science, Nanyang Technological University, 50 Nanyang Ave, 639798, Singapore, Singapore. gumile@ntu.edu.sg., Kwek LC; Quantum Science and Engineering Centre (QSec), Nanyang Technological University, Singapore, 639798, Singapore. cqtklc@nus.edu.sg.; Centre for Quantum Technologies, National University of Singapore, Singapore, 117543, Singapore. cqtklc@nus.edu.sg., Liu AQ; Quantum Science and Engineering Centre (QSec), Nanyang Technological University, Singapore, 639798, Singapore. eaqliu@ntu.edu.sg.
Jazyk: angličtina
Zdroj: Nature communications [Nat Commun] 2022 Feb 24; Vol. 13 (1), pp. 1044. Date of Electronic Publication: 2022 Feb 24.
DOI: 10.1038/s41467-022-28702-0
Abstrakt: Large-scale, highly integrated and low-power-consuming hardware is becoming progressively more important for realizing optical neural networks (ONNs) capable of advanced optical computing. Traditional experimental implementations need N 2 units such as Mach-Zehnder interferometers (MZIs) for an input dimension N to realize typical computing operations (convolutions and matrix multiplication), resulting in limited scalability and consuming excessive power. Here, we propose the integrated diffractive optical network for implementing parallel Fourier transforms, convolution operations and application-specific optical computing using two ultracompact diffractive cells (Fourier transform operation) and only N MZIs. The footprint and energy consumption scales linearly with the input data dimension, instead of the quadratic scaling in the traditional ONN framework. A ~10-fold reduction in both footprint and energy consumption, as well as equal high accuracy with previous MZI-based ONNs was experimentally achieved for computations performed on the MNIST and Fashion-MNIST datasets. The integrated diffractive optical network (IDNN) chip demonstrates a promising avenue towards scalable and low-power-consumption optical computational chips for optical-artificial-intelligence.
(© 2022. The Author(s).)
Databáze: MEDLINE
Externí odkaz: