Implementing a deep neural network using a single nonlinear element and delayed feedback

André Röhm; Florian Stelzer; Tomoaki Niiyama; Raul Vicente; Ingo Fischer; Makoto Naruse; Serhiy Yanchuk; Satoshi Sunada

doi:10.1364/NLO.2023.M4A.26

Implementing a deep neural network using a single nonlinear element and delayed feedback

André Röhm
, Florian Stelzer
, Tomoaki Niiyama
, Raul Vicente
, Ingo Fischer
, Makoto Naruse
, Serhiy Yanchuk
, Satoshi Sunada

Research output: Contribution to conference › Paper › peer-review

Abstract

Optically implemented neural networks promise faster speeds and lower costs. However, large size networks remain challenging. We derive how to emulate a deep neural network with just a single nonlinear element using delayed feedback signals.

Original language	English
DOIs	https://doi.org/10.1364/NLO.2023.M4A.26
Publication status	Published - 2023
Externally published	Yes
Event	2023 Optica Nonlinear Optics Topical Meeting, NLO 2023 - Honolulu, United States Duration: 10 Jul 2023 → 13 Jul 2023

Conference

Conference	2023 Optica Nonlinear Optics Topical Meeting, NLO 2023
Country/Territory	United States
City	Honolulu
Period	10/07/23 → 13/07/23

Access to Document

10.1364/NLO.2023.M4A.26

Cite this

@conference{77dbe28bea304c02bd7ac5382d247440,

title = "Implementing a deep neural network using a single nonlinear element and delayed feedback",

abstract = "Optically implemented neural networks promise faster speeds and lower costs. However, large size networks remain challenging. We derive how to emulate a deep neural network with just a single nonlinear element using delayed feedback signals.",

author = "Andr{\'e} R{\"o}hm and Florian Stelzer and Tomoaki Niiyama and Raul Vicente and Ingo Fischer and Makoto Naruse and Serhiy Yanchuk and Satoshi Sunada",

note = "Publisher Copyright: Optica Nonlinear Optics Topical Meeting 2023 {\textcopyright} Optica Publishing Group 2023 {\textcopyright} 2023 The Author(s); 2023 Optica Nonlinear Optics Topical Meeting, NLO 2023 ; Conference date: 10-07-2023 Through 13-07-2023",

year = "2023",

doi = "10.1364/NLO.2023.M4A.26",

language = "English",

}

TY - CONF

T1 - Implementing a deep neural network using a single nonlinear element and delayed feedback

AU - Röhm, André

AU - Stelzer, Florian

AU - Niiyama, Tomoaki

AU - Vicente, Raul

AU - Fischer, Ingo

AU - Naruse, Makoto

AU - Yanchuk, Serhiy

AU - Sunada, Satoshi

PY - 2023

Y1 - 2023

N2 - Optically implemented neural networks promise faster speeds and lower costs. However, large size networks remain challenging. We derive how to emulate a deep neural network with just a single nonlinear element using delayed feedback signals.

AB - Optically implemented neural networks promise faster speeds and lower costs. However, large size networks remain challenging. We derive how to emulate a deep neural network with just a single nonlinear element using delayed feedback signals.

UR - https://www.scopus.com/pages/publications/85192519098

U2 - 10.1364/NLO.2023.M4A.26

DO - 10.1364/NLO.2023.M4A.26

M3 - Paper

AN - SCOPUS:85192519098

T2 - 2023 Optica Nonlinear Optics Topical Meeting, NLO 2023

Y2 - 10 July 2023 through 13 July 2023

ER -

Implementing a deep neural network using a single nonlinear element and delayed feedback

Abstract

Conference

Access to Document

Other files and links

Fingerprint

Cite this