Equalization performance and complexity analysis of dynamic deep neural networks in long haul transmission systems

Oleg Sidelnikov, Alexey Redyuk, Stylianos Sygletos

Research output: Contribution to journalArticle

Abstract

We investigate the application of dynamic deep neural networks for nonlinear equalization in long haul transmission systems. Through extensive numerical analysis we identify their optimum dimensions and calculate their computational complexity as a function of system length. Performing comparison with traditional back-propagation based nonlinear compensation of 2 steps-per-span and 2 samples-per-symbol, we demonstrate equivalent mitigation performance at significantly lower computational cost.
Original languageEnglish
Pages (from-to)32765-32776
JournalOptics Express
Volume26
Issue number25
Early online date29 Nov 2018
DOIs
Publication statusPublished - 10 Dec 2018

Fingerprint

numerical analysis
costs

Bibliographical note

An OSA-formatted open access journal article PDF may be governed by the OSA Open Access Publishing Agreement signed by the author and any applicable copyright laws. Authors and readers may use, reuse, and build upon the article, or use it for text or data mining without asking prior permission from the publisher or the Author(s), as long as the purpose is non-commercial and appropriate attribution is maintained.

Cite this

@article{af543765af2049dba83adf1304903331,
title = "Equalization performance and complexity analysis of dynamic deep neural networks in long haul transmission systems",
abstract = "We investigate the application of dynamic deep neural networks for nonlinear equalization in long haul transmission systems. Through extensive numerical analysis we identify their optimum dimensions and calculate their computational complexity as a function of system length. Performing comparison with traditional back-propagation based nonlinear compensation of 2 steps-per-span and 2 samples-per-symbol, we demonstrate equivalent mitigation performance at significantly lower computational cost.",
author = "Oleg Sidelnikov and Alexey Redyuk and Stylianos Sygletos",
note = "An OSA-formatted open access journal article PDF may be governed by the OSA Open Access Publishing Agreement signed by the author and any applicable copyright laws. Authors and readers may use, reuse, and build upon the article, or use it for text or data mining without asking prior permission from the publisher or the Author(s), as long as the purpose is non-commercial and appropriate attribution is maintained.",
year = "2018",
month = "12",
day = "10",
doi = "10.1364/OE.26.032765",
language = "English",
volume = "26",
pages = "32765--32776",
journal = "Optics Express",
issn = "1094-4087",
publisher = "The Optical Society",
number = "25",

}

Equalization performance and complexity analysis of dynamic deep neural networks in long haul transmission systems. / Sidelnikov, Oleg; Redyuk, Alexey; Sygletos, Stylianos.

In: Optics Express, Vol. 26, No. 25, 10.12.2018, p. 32765-32776.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Equalization performance and complexity analysis of dynamic deep neural networks in long haul transmission systems

AU - Sidelnikov, Oleg

AU - Redyuk, Alexey

AU - Sygletos, Stylianos

N1 - An OSA-formatted open access journal article PDF may be governed by the OSA Open Access Publishing Agreement signed by the author and any applicable copyright laws. Authors and readers may use, reuse, and build upon the article, or use it for text or data mining without asking prior permission from the publisher or the Author(s), as long as the purpose is non-commercial and appropriate attribution is maintained.

PY - 2018/12/10

Y1 - 2018/12/10

N2 - We investigate the application of dynamic deep neural networks for nonlinear equalization in long haul transmission systems. Through extensive numerical analysis we identify their optimum dimensions and calculate their computational complexity as a function of system length. Performing comparison with traditional back-propagation based nonlinear compensation of 2 steps-per-span and 2 samples-per-symbol, we demonstrate equivalent mitigation performance at significantly lower computational cost.

AB - We investigate the application of dynamic deep neural networks for nonlinear equalization in long haul transmission systems. Through extensive numerical analysis we identify their optimum dimensions and calculate their computational complexity as a function of system length. Performing comparison with traditional back-propagation based nonlinear compensation of 2 steps-per-span and 2 samples-per-symbol, we demonstrate equivalent mitigation performance at significantly lower computational cost.

UR - https://www.osapublishing.org/abstract.cfm?URI=oe-26-25-32765

U2 - 10.1364/OE.26.032765

DO - 10.1364/OE.26.032765

M3 - Article

VL - 26

SP - 32765

EP - 32776

JO - Optics Express

JF - Optics Express

SN - 1094-4087

IS - 25

ER -