Optical Versus Electronic Implementation of Probabilistic Graphical Inference and Experimental Device Demonstration Using Nonlinear Photonics

Masoud Babaeian, Patrick Keiffer, Mark A Neifeld, Ratchaneekorn Thamvichai, Robert A Norwood, Pierre Alexandre Blanche, John W Wissinger, Nasser N Peyghambarian

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

The probabilistic inference model has been widely used in various areas, such as error-control coding, machine learning, speech recognition, artificial intelligence, and statistics. In this paper, we study both computation and communications power consumption of optical-based and electronic-based implementations of the probabilistic inference algorithm used in solving large scale problems. Our analysis indicates that the optical implementation provides substantial reduction for power and area compare to the electronic-based solutions as problems become large. For a network with 1 million nodes and 100 alphabet size, our proposed wavelength multiplexed all-optical implementation requires approximately 200 kilowatts (kW) of power as compared with 1.47 gigawatts (GW) and 1.7 megawatts (MW) using CPU-based and subthreshold VLSI-based systems, respectively. The optical-based solution is tolerant to shot noise and imperfections of optical modules used in the architecture as well. We also performed an all-optical experimental verification of a graphical inference as the proof of concept and have demonstrated the essential mathematical operations, multiplication, and normalization (division), in photonics operations using nonlinear bulk materials. The normalization and multiplication are shown optically through a pump-probe saturation process and a logarithm-summation-exponential (log-sum-exp) operation, respectively. We used single mode silicon waveguide and single-wall carbon nanotube (SWCNT) as nonlinear optical materials to implement logarithm and exponential operations, respectively. The SWCNT is also used as the nonlinear component in the pump-probe saturation experiment to implement the normalization function.

Original languageEnglish (US)
Article number8471109
JournalIEEE Photonics Journal
Volume10
Issue number5
DOIs
StatePublished - Oct 1 2018

Fingerprint

inference
Photonics
Carbon nanotubes
Demonstrations
Pumps
photonics
Shot noise
Optical materials
logarithms
Speech recognition
electronics
multiplication
Artificial intelligence
Program processors
Learning systems
Waveguides
Electric power utilization
carbon nanotubes
Statistics
pumps

Keywords

  • nonlinear optical devices
  • Nonlinear optics
  • optical computing
  • photonics
  • ultrafast optics.

ASJC Scopus subject areas

  • Atomic and Molecular Physics, and Optics
  • Electrical and Electronic Engineering

Cite this

@article{cba44116a5d541d78fb6c1bdcf7c8fc3,
title = "Optical Versus Electronic Implementation of Probabilistic Graphical Inference and Experimental Device Demonstration Using Nonlinear Photonics",
abstract = "The probabilistic inference model has been widely used in various areas, such as error-control coding, machine learning, speech recognition, artificial intelligence, and statistics. In this paper, we study both computation and communications power consumption of optical-based and electronic-based implementations of the probabilistic inference algorithm used in solving large scale problems. Our analysis indicates that the optical implementation provides substantial reduction for power and area compare to the electronic-based solutions as problems become large. For a network with 1 million nodes and 100 alphabet size, our proposed wavelength multiplexed all-optical implementation requires approximately 200 kilowatts (kW) of power as compared with 1.47 gigawatts (GW) and 1.7 megawatts (MW) using CPU-based and subthreshold VLSI-based systems, respectively. The optical-based solution is tolerant to shot noise and imperfections of optical modules used in the architecture as well. We also performed an all-optical experimental verification of a graphical inference as the proof of concept and have demonstrated the essential mathematical operations, multiplication, and normalization (division), in photonics operations using nonlinear bulk materials. The normalization and multiplication are shown optically through a pump-probe saturation process and a logarithm-summation-exponential (log-sum-exp) operation, respectively. We used single mode silicon waveguide and single-wall carbon nanotube (SWCNT) as nonlinear optical materials to implement logarithm and exponential operations, respectively. The SWCNT is also used as the nonlinear component in the pump-probe saturation experiment to implement the normalization function.",
keywords = "nonlinear optical devices, Nonlinear optics, optical computing, photonics, ultrafast optics.",
author = "Masoud Babaeian and Patrick Keiffer and Neifeld, {Mark A} and Ratchaneekorn Thamvichai and Norwood, {Robert A} and Blanche, {Pierre Alexandre} and Wissinger, {John W} and Peyghambarian, {Nasser N}",
year = "2018",
month = "10",
day = "1",
doi = "10.1109/JPHOT.2018.2871822",
language = "English (US)",
volume = "10",
journal = "IEEE Photonics Journal",
issn = "1943-0655",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "5",

}

TY - JOUR

T1 - Optical Versus Electronic Implementation of Probabilistic Graphical Inference and Experimental Device Demonstration Using Nonlinear Photonics

AU - Babaeian, Masoud

AU - Keiffer, Patrick

AU - Neifeld, Mark A

AU - Thamvichai, Ratchaneekorn

AU - Norwood, Robert A

AU - Blanche, Pierre Alexandre

AU - Wissinger, John W

AU - Peyghambarian, Nasser N

PY - 2018/10/1

Y1 - 2018/10/1

N2 - The probabilistic inference model has been widely used in various areas, such as error-control coding, machine learning, speech recognition, artificial intelligence, and statistics. In this paper, we study both computation and communications power consumption of optical-based and electronic-based implementations of the probabilistic inference algorithm used in solving large scale problems. Our analysis indicates that the optical implementation provides substantial reduction for power and area compare to the electronic-based solutions as problems become large. For a network with 1 million nodes and 100 alphabet size, our proposed wavelength multiplexed all-optical implementation requires approximately 200 kilowatts (kW) of power as compared with 1.47 gigawatts (GW) and 1.7 megawatts (MW) using CPU-based and subthreshold VLSI-based systems, respectively. The optical-based solution is tolerant to shot noise and imperfections of optical modules used in the architecture as well. We also performed an all-optical experimental verification of a graphical inference as the proof of concept and have demonstrated the essential mathematical operations, multiplication, and normalization (division), in photonics operations using nonlinear bulk materials. The normalization and multiplication are shown optically through a pump-probe saturation process and a logarithm-summation-exponential (log-sum-exp) operation, respectively. We used single mode silicon waveguide and single-wall carbon nanotube (SWCNT) as nonlinear optical materials to implement logarithm and exponential operations, respectively. The SWCNT is also used as the nonlinear component in the pump-probe saturation experiment to implement the normalization function.

AB - The probabilistic inference model has been widely used in various areas, such as error-control coding, machine learning, speech recognition, artificial intelligence, and statistics. In this paper, we study both computation and communications power consumption of optical-based and electronic-based implementations of the probabilistic inference algorithm used in solving large scale problems. Our analysis indicates that the optical implementation provides substantial reduction for power and area compare to the electronic-based solutions as problems become large. For a network with 1 million nodes and 100 alphabet size, our proposed wavelength multiplexed all-optical implementation requires approximately 200 kilowatts (kW) of power as compared with 1.47 gigawatts (GW) and 1.7 megawatts (MW) using CPU-based and subthreshold VLSI-based systems, respectively. The optical-based solution is tolerant to shot noise and imperfections of optical modules used in the architecture as well. We also performed an all-optical experimental verification of a graphical inference as the proof of concept and have demonstrated the essential mathematical operations, multiplication, and normalization (division), in photonics operations using nonlinear bulk materials. The normalization and multiplication are shown optically through a pump-probe saturation process and a logarithm-summation-exponential (log-sum-exp) operation, respectively. We used single mode silicon waveguide and single-wall carbon nanotube (SWCNT) as nonlinear optical materials to implement logarithm and exponential operations, respectively. The SWCNT is also used as the nonlinear component in the pump-probe saturation experiment to implement the normalization function.

KW - nonlinear optical devices

KW - Nonlinear optics

KW - optical computing

KW - photonics

KW - ultrafast optics.

UR - http://www.scopus.com/inward/record.url?scp=85054236027&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85054236027&partnerID=8YFLogxK

U2 - 10.1109/JPHOT.2018.2871822

DO - 10.1109/JPHOT.2018.2871822

M3 - Article

AN - SCOPUS:85054236027

VL - 10

JO - IEEE Photonics Journal

JF - IEEE Photonics Journal

SN - 1943-0655

IS - 5

M1 - 8471109

ER -