Photonic machine learning with on-chip diffractive optics

Tingzhao Fu; Yubin Zang; Yuyao Huang; Zhenmin Du; Honghao Huang; Chengyang Hu; Minghua Chen; Sigang Yang; Hongwei Chen

doi:10.1038/s41467-022-35772-7

Nature Communications (Jan 2023)

Photonic machine learning with on-chip diffractive optics

Tingzhao Fu,
Yubin Zang,
Yuyao Huang,
Zhenmin Du,
Honghao Huang,
Chengyang Hu,
Minghua Chen,
Sigang Yang,
Hongwei Chen

Affiliations

Tingzhao Fu: Beijing National Research Center for Information Science and Technology, Department of Electronic Engineering, Tsinghua University
Yubin Zang: Beijing National Research Center for Information Science and Technology, Department of Electronic Engineering, Tsinghua University
Yuyao Huang: Beijing National Research Center for Information Science and Technology, Department of Electronic Engineering, Tsinghua University
Zhenmin Du: Beijing National Research Center for Information Science and Technology, Department of Electronic Engineering, Tsinghua University
Honghao Huang: Beijing National Research Center for Information Science and Technology, Department of Electronic Engineering, Tsinghua University
Chengyang Hu: Beijing National Research Center for Information Science and Technology, Department of Electronic Engineering, Tsinghua University
Minghua Chen: Beijing National Research Center for Information Science and Technology, Department of Electronic Engineering, Tsinghua University
Sigang Yang: Beijing National Research Center for Information Science and Technology, Department of Electronic Engineering, Tsinghua University
Hongwei Chen: Beijing National Research Center for Information Science and Technology, Department of Electronic Engineering, Tsinghua University

DOI: https://doi.org/10.1038/s41467-022-35772-7
Journal volume & issue: Vol. 14, no. 1
pp. 1 – 10

Abstract

Read online

Integrating diffractive optical neural networks (DONN) would reduce errors due to bulky components and calibration. Here, the authors exploit integrated 1D dielectric metasurfaces to realise an on-chip DONN device with 90% classification accuracy, computing at 10^16 flops/mm^2 and consuming 10E-17 J/Flop.

Published in Nature Communications

ISSN: 2041-1723 (Online)
Publisher: Nature Portfolio
Country of publisher: United Kingdom
LCC subjects: Science
Website: https://www.nature.com/ncomms/

About the journal