Comparison of Models for Training Optical Matrix Multipliers in
  Neuromorphic PICs

Ali Cem; Siqi Yan; Uiara Celine de Moura; Yunhong Ding; Darko Zibar; Francesco Da Ros

Comparison of Models for Training Optical Matrix Multipliers in Neuromorphic PICs

Ali Cem^*, Siqi Yan, Uiara Celine de Moura, Yunhong Ding, Darko Zibar, Francesco Da Ros

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

48 Downloads (Pure)

Abstract

We experimentally compare simple physics-based vs. data-driven neural-network-based models for offline training of programmable photonic chips using Mach-Zehnder interferometer meshes. The neural-network model outperforms physics-based models for a chip with thermal crosstalk, yielding increased testing accuracy.

Original language	English
Title of host publication	Proceedings of OFC 2022
Number of pages	3
Publisher	Optical Society of America (OSA)
Publication date	2022
Article number	M2G.5
Publication status	Published - 2022
Event	2022 Optical Fiber Communications Conference and Exhibition - San Diego Convention Center, San Diego, United States Duration: 6 Mar 2022 → 10 Mar 2022

Conference

Conference	2022 Optical Fiber Communications Conference and Exhibition
Location	San Diego Convention Center
Country/Territory	United States
City	San Diego
Period	06/03/2022 → 10/03/2022

Access to Document

FulltextAccepted author manuscript, 3.37 MB

OpenUrl availability

Full text

Cite this

@inproceedings{0ecae6a1cd6b4874a87178bf60f7db03,

title = "Comparison of Models for Training Optical Matrix Multipliers in Neuromorphic PICs",

abstract = "We experimentally compare simple physics-based vs. data-driven neural-network-based models for offline training of programmable photonic chips using Mach-Zehnder interferometer meshes. The neural-network model outperforms physics-based models for a chip with thermal crosstalk, yielding increased testing accuracy.",

author = "Ali Cem and Siqi Yan and {de Moura}, {Uiara Celine} and Yunhong Ding and Darko Zibar and {Da Ros}, Francesco",

year = "2022",

language = "English",

booktitle = "Proceedings of OFC 2022",

publisher = "Optical Society of America (OSA)",

note = "2022 Optical Fiber Communications Conference and Exhibition, OFC 2022 ; Conference date: 06-03-2022 Through 10-03-2022",

}

TY - GEN

T1 - Comparison of Models for Training Optical Matrix Multipliers in Neuromorphic PICs

AU - Cem, Ali

AU - Yan, Siqi

AU - de Moura, Uiara Celine

AU - Ding, Yunhong

AU - Zibar, Darko

AU - Da Ros, Francesco

PY - 2022

Y1 - 2022

N2 - We experimentally compare simple physics-based vs. data-driven neural-network-based models for offline training of programmable photonic chips using Mach-Zehnder interferometer meshes. The neural-network model outperforms physics-based models for a chip with thermal crosstalk, yielding increased testing accuracy.

AB - We experimentally compare simple physics-based vs. data-driven neural-network-based models for offline training of programmable photonic chips using Mach-Zehnder interferometer meshes. The neural-network model outperforms physics-based models for a chip with thermal crosstalk, yielding increased testing accuracy.

M3 - Article in proceedings

BT - Proceedings of OFC 2022

PB - Optical Society of America (OSA)

T2 - 2022 Optical Fiber Communications Conference and Exhibition

Y2 - 6 March 2022 through 10 March 2022

ER -