Noise-Resistant Optical Implementation of Analogue Neural Networks

Diego Arguello Ron; Morteza Kamalian Kopae; Sergei K. Turitsyn

doi:10.1109/CLEO/Europe-EQEC52157.2021.9541571

Noise-Resistant Optical Implementation of Analogue Neural Networks

Diego Arguello Ron, Morteza Kamalian Kopae, Sergei K. Turitsyn

Research output: Chapter in Book/Published conference output › Conference publication

Abstract

Analogue artificial neural networks are widely considered as promising computational models that more closely imitate the information processing capabilities of the human brain compared to digital neural networks. The significant computation power and the much reduced power consumption per operation make the analogue implementation of neural networks very attractive. There is an active research on artificial neural networks (ANNs) implementation using both analogue photonic and electronic hardware [1] – [4] . However, compared to digital realisations the conventional analogue systems are more sensitive to the noise that is inevitably present in practical implementations [2] , [3] . Noise properties in ANNs have been studied both in the electronic and photonic domains. However, photonic ANNs are much less investigated compared to the electronic implementations, for which some training techniques have been proposed to enhance ANNs robustness against noise [1] , [4] .

Original language	English
Title of host publication	2021 Conference on Lasers and Electro-Optics Europe and European Quantum Electronics Conference, CLEO/Europe-EQEC 2021
Publisher	IEEE
ISBN (Electronic)	978-1-6654-1876-8
ISBN (Print)	978-1-6654-4804-8
DOIs	https://doi.org/10.1109/CLEO/Europe-EQEC52157.2021.9541571
Publication status	Published - 30 Sept 2021
Event	2021 Conference on Lasers and Electro-Optics Europe & European Quantum Electronics Conference (CLEO/Europe-EQEC) - Munich, Germany Duration: 21 Jun 2021 → 25 Jun 2021

Conference

Conference	2021 Conference on Lasers and Electro-Optics Europe & European Quantum Electronics Conference (CLEO/Europe-EQEC)
Country/Territory	Germany
City	Munich
Period	21/06/21 → 25/06/21

Bibliographical note

© 2021 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.

Keywords

Training
Power demand
Artificial neural networks
Optical computing
Optical fiber networks
Robustness
Optical sensors

Access to Document

10.1109/CLEO/Europe-EQEC52157.2021.9541571

Noise-Resistant Optical Implementation of Analogue Neural Networks
© 2021 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.
Accepted author manuscript, 73.2 KB

Cite this

@inproceedings{50f191e3d6164d2aba8effcd041bb6d1,

title = "Noise-Resistant Optical Implementation of Analogue Neural Networks",

abstract = "Analogue artificial neural networks are widely considered as promising computational models that more closely imitate the information processing capabilities of the human brain compared to digital neural networks. The significant computation power and the much reduced power consumption per operation make the analogue implementation of neural networks very attractive. There is an active research on artificial neural networks (ANNs) implementation using both analogue photonic and electronic hardware [1] – [4] . However, compared to digital realisations the conventional analogue systems are more sensitive to the noise that is inevitably present in practical implementations [2] , [3] . Noise properties in ANNs have been studied both in the electronic and photonic domains. However, photonic ANNs are much less investigated compared to the electronic implementations, for which some training techniques have been proposed to enhance ANNs robustness against noise [1] , [4] .",

keywords = "Training, Power demand, Artificial neural networks, Optical computing, Optical fiber networks, Robustness, Optical sensors",

author = "{Arguello Ron}, Diego and {Kamalian Kopae}, Morteza and Turitsyn, {Sergei K.}",

note = "{\textcopyright} 2021 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.; 2021 Conference on Lasers and Electro-Optics Europe & European Quantum Electronics Conference (CLEO/Europe-EQEC) ; Conference date: 21-06-2021 Through 25-06-2021",

year = "2021",

month = sep,

day = "30",

doi = "10.1109/CLEO/Europe-EQEC52157.2021.9541571",

language = "English",

isbn = "978-1-6654-4804-8",

booktitle = "2021 Conference on Lasers and Electro-Optics Europe and European Quantum Electronics Conference, CLEO/Europe-EQEC 2021",

publisher = "IEEE",

address = "United States",

}

Arguello Ron, D, Kamalian Kopae, M & Turitsyn, SK 2021, Noise-Resistant Optical Implementation of Analogue Neural Networks. in 2021 Conference on Lasers and Electro-Optics Europe and European Quantum Electronics Conference, CLEO/Europe-EQEC 2021. IEEE, 2021 Conference on Lasers and Electro-Optics Europe & European Quantum Electronics Conference (CLEO/Europe-EQEC), Munich, Germany, 21/06/21. https://doi.org/10.1109/CLEO/Europe-EQEC52157.2021.9541571

TY - GEN

T1 - Noise-Resistant Optical Implementation of Analogue Neural Networks

AU - Arguello Ron, Diego

AU - Kamalian Kopae, Morteza

AU - Turitsyn, Sergei K.

N1 - © 2021 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.

PY - 2021/9/30

Y1 - 2021/9/30

N2 - Analogue artificial neural networks are widely considered as promising computational models that more closely imitate the information processing capabilities of the human brain compared to digital neural networks. The significant computation power and the much reduced power consumption per operation make the analogue implementation of neural networks very attractive. There is an active research on artificial neural networks (ANNs) implementation using both analogue photonic and electronic hardware [1] – [4] . However, compared to digital realisations the conventional analogue systems are more sensitive to the noise that is inevitably present in practical implementations [2] , [3] . Noise properties in ANNs have been studied both in the electronic and photonic domains. However, photonic ANNs are much less investigated compared to the electronic implementations, for which some training techniques have been proposed to enhance ANNs robustness against noise [1] , [4] .

AB - Analogue artificial neural networks are widely considered as promising computational models that more closely imitate the information processing capabilities of the human brain compared to digital neural networks. The significant computation power and the much reduced power consumption per operation make the analogue implementation of neural networks very attractive. There is an active research on artificial neural networks (ANNs) implementation using both analogue photonic and electronic hardware [1] – [4] . However, compared to digital realisations the conventional analogue systems are more sensitive to the noise that is inevitably present in practical implementations [2] , [3] . Noise properties in ANNs have been studied both in the electronic and photonic domains. However, photonic ANNs are much less investigated compared to the electronic implementations, for which some training techniques have been proposed to enhance ANNs robustness against noise [1] , [4] .

KW - Training

KW - Power demand

KW - Artificial neural networks

KW - Optical computing

KW - Optical fiber networks

KW - Robustness

KW - Optical sensors

UR - https://ieeexplore.ieee.org/document/9541571

UR - https://www.cleoeurope.org/

UR - http://www.scopus.com/inward/record.url?scp=85117585393&partnerID=8YFLogxK

U2 - 10.1109/CLEO/Europe-EQEC52157.2021.9541571

DO - 10.1109/CLEO/Europe-EQEC52157.2021.9541571

M3 - Conference publication

SN - 978-1-6654-4804-8

BT - 2021 Conference on Lasers and Electro-Optics Europe and European Quantum Electronics Conference, CLEO/Europe-EQEC 2021

PB - IEEE

T2 - 2021 Conference on Lasers and Electro-Optics Europe & European Quantum Electronics Conference (CLEO/Europe-EQEC)

Y2 - 21 June 2021 through 25 June 2021

ER -

Noise-Resistant Optical Implementation of Analogue Neural Networks

Abstract

Conference

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this