Skip to main navigation Skip to search Skip to main content

Hyper-multiplexed, Ultralow-Energy Optical Neural Networks on Thin-Film Lithium Niobate

  • Shaoyuan Ou*
  • , Alexander Sludds
  • , Ryan Hamerly
  • , Eric Zhong
  • , Ke Zhang
  • , Hanke Feng
  • , Cheng Wang
  • , Dirk Englund
  • , Mengjie Yu
  • , Zaijun Chen*
  • *Corresponding author for this work

Research output: Chapters, Conference Papers, Creative and Literary WorksRGC 32 - Refereed conference paper (with host publication)peer-review

Abstract

We demonstrate a large-scale wavelength-time-space-multiplexed optical neural network using high-bandwidth (>40 GHz) electro-optic modulators at CMOS-compatible voltages (Vπ=1.3 V). Parallel computing with 7 wavelengths (over 1-THz) achieves 6-bit precision for accurate image classification. © 2024 The Author(s) © Optica Publishing Group 2024
Original languageEnglish
Title of host publicationCLEO: Science and Innovations 2024
PublisherOptical Society of America
ISBN (Electronic)978-1-957171-39-5
DOIs
Publication statusPublished - May 2024
Event2024 Conference on Lasers and Electro-Optics (CLEO 2024) - Charlotte, United States
Duration: 5 May 202410 May 2024

Publication series

NameCLEO: Science and Innovations, CLEO: S and I in Proceedings CLEO, Part of Conference on Lasers and Electro-Optics

Conference

Conference2024 Conference on Lasers and Electro-Optics (CLEO 2024)
PlaceUnited States
CityCharlotte
Period5/05/2410/05/24

Fingerprint

Dive into the research topics of 'Hyper-multiplexed, Ultralow-Energy Optical Neural Networks on Thin-Film Lithium Niobate'. Together they form a unique fingerprint.

Cite this