Energy-Efficient Neural Network Inference with Microcavity Exciton Polaritons

M. Matuszewski, A. Opala, R. Mirek, M. Furman, M. Król, K. Tyszka, T.C.H. Liew, D. Ballarini, D. Sanvitto, J. Szczytko, and B. Piętka
Phys. Rev. Applied 16, 024045 – Published 25 August 2021

Abstract

We propose all-optical neural networks characterized by very high energy efficiency and performance density of inference. We argue that the use of microcavity exciton polaritons allows one to take advantage of the properties of both photons and electrons in a seamless manner. This results in strong optical nonlinearity without the use of optoelectronic conversion. We propose a design of a realistic neural network and estimate energy cost to be at the level of attojoules per bit, also when including the optoelectronic conversion at the input and output of the network, several orders of magnitude below state-of-the-art hardware implementations. We propose two kinds of nonlinear binarized nodes based either on optical phase shifts and interferometry or on polariton spin rotations.

  • Figure
  • Figure
  • Figure
  • Figure
  • Figure
  • Figure
  • Figure
  • Received 31 May 2021
  • Revised 21 July 2021
  • Accepted 29 July 2021
  • Corrected 8 September 2021

DOI:https://doi.org/10.1103/PhysRevApplied.16.024045

© 2021 American Physical Society

Physics Subject Headings (PhySH)

Condensed Matter, Materials & Applied Physics

Corrections

8 September 2021

Correction: The omission of additional support statements in the Acknowledgment section has been fixed.

Authors & Affiliations

M. Matuszewski1,*, A. Opala1, R. Mirek2, M. Furman2, M. Król2, K. Tyszka2, T.C.H. Liew3, D. Ballarini4, D. Sanvitto4,5, J. Szczytko2, and B. Piętka2

  • 1Institute of Physics, Polish Academy of Sciences, Al. Lotników 32/46, Warsaw PL-02-668, Poland
  • 2Institute of Experimental Physics, Faculty of Physics, University of Warsaw, ul. Pasteura 5, Warsaw PL-02-093, Poland
  • 3Division of Physics and Applied Physics, Nanyang Technological University 637371, Singapore
  • 4CNR NANOTEC-Institute of Nanotechnology, Via Monteroni, Lecce 73100, Italy
  • 5INFN, Sez. Lecce, Lecce 73100, Italy

  • *mmatu@ifpan.edu.pl

Article Text (Subscription Required)

Click to Expand

References (Subscription Required)

Click to Expand
Issue

Vol. 16, Iss. 2 — August 2021

Subject Areas
Reuse & Permissions
Access Options
Author publication services for translation and copyediting assistance advertisement

Authorization Required


×
×

Images

×

Sign up to receive regular email alerts from Physical Review Applied

Log In

Cancel
×

Search


Article Lookup

Paste a citation or DOI

Enter a citation
×