Nonlinear optical components for all-optical probabilistic graphical model

Masoud Babaeian, Pierre Alexandre Blanche, Robert A Norwood, Tommi Kaplas, Patrick Keiffer, Yuri Svirko, Taylor G. Allen, Vincent W. Chen, San Hui Chi, Joseph W. Perry, Seth R. Marder, Mark A Neifeld, Nasser N Peyghambarian

Research output: Contribution to journalArticle

6 Citations (Scopus)

Abstract

The probabilistic graphical models (PGMs) are tools that are used to compute probability distributions over large and complex interacting variables. They have applications in social networks, speech recognition, artificial intelligence, machine learning, and many more areas. Here, we present an all-optical implementation of a PGM through the sum-product message passing algorithm (SPMPA) governed by a wavelength multiplexing architecture. As a proof-of-concept, we demonstrate the use of optics to solve a two node graphical model governed by SPMPA and successfully map the message passing algorithm onto photonics operations. The essential mathematical functions required for this algorithm, including multiplication and division, are implemented using nonlinear optics in thin film materials. The multiplication and division are demonstrated through a logarithm-summation-exponentiation operation and a pump-probe saturation process, respectively. The fundamental bottlenecks for the scalability of the presented scheme are discussed as well.

Original languageEnglish (US)
Article number2128
JournalNature Communications
Volume9
Issue number1
DOIs
StatePublished - Dec 1 2018

Fingerprint

Optical Devices
Statistical Models
Message passing
messages
multiplication
division
Optics and Photonics
artificial intelligence
Nonlinear optics
machine learning
Artificial Intelligence
nonlinear optics
speech recognition
products
logarithms
Speech recognition
multiplexing
Multiplexing
Social Support
Photonics

ASJC Scopus subject areas

  • Chemistry(all)
  • Biochemistry, Genetics and Molecular Biology(all)
  • Physics and Astronomy(all)

Cite this

Nonlinear optical components for all-optical probabilistic graphical model. / Babaeian, Masoud; Blanche, Pierre Alexandre; Norwood, Robert A; Kaplas, Tommi; Keiffer, Patrick; Svirko, Yuri; Allen, Taylor G.; Chen, Vincent W.; Chi, San Hui; Perry, Joseph W.; Marder, Seth R.; Neifeld, Mark A; Peyghambarian, Nasser N.

In: Nature Communications, Vol. 9, No. 1, 2128, 01.12.2018.

Research output: Contribution to journalArticle

Babaeian, M, Blanche, PA, Norwood, RA, Kaplas, T, Keiffer, P, Svirko, Y, Allen, TG, Chen, VW, Chi, SH, Perry, JW, Marder, SR, Neifeld, MA & Peyghambarian, NN 2018, 'Nonlinear optical components for all-optical probabilistic graphical model', Nature Communications, vol. 9, no. 1, 2128. https://doi.org/10.1038/s41467-018-04578-x
Babaeian, Masoud ; Blanche, Pierre Alexandre ; Norwood, Robert A ; Kaplas, Tommi ; Keiffer, Patrick ; Svirko, Yuri ; Allen, Taylor G. ; Chen, Vincent W. ; Chi, San Hui ; Perry, Joseph W. ; Marder, Seth R. ; Neifeld, Mark A ; Peyghambarian, Nasser N. / Nonlinear optical components for all-optical probabilistic graphical model. In: Nature Communications. 2018 ; Vol. 9, No. 1.
@article{285582dc82534beb93a8a77ab9f1fc33,
title = "Nonlinear optical components for all-optical probabilistic graphical model",
abstract = "The probabilistic graphical models (PGMs) are tools that are used to compute probability distributions over large and complex interacting variables. They have applications in social networks, speech recognition, artificial intelligence, machine learning, and many more areas. Here, we present an all-optical implementation of a PGM through the sum-product message passing algorithm (SPMPA) governed by a wavelength multiplexing architecture. As a proof-of-concept, we demonstrate the use of optics to solve a two node graphical model governed by SPMPA and successfully map the message passing algorithm onto photonics operations. The essential mathematical functions required for this algorithm, including multiplication and division, are implemented using nonlinear optics in thin film materials. The multiplication and division are demonstrated through a logarithm-summation-exponentiation operation and a pump-probe saturation process, respectively. The fundamental bottlenecks for the scalability of the presented scheme are discussed as well.",
author = "Masoud Babaeian and Blanche, {Pierre Alexandre} and Norwood, {Robert A} and Tommi Kaplas and Patrick Keiffer and Yuri Svirko and Allen, {Taylor G.} and Chen, {Vincent W.} and Chi, {San Hui} and Perry, {Joseph W.} and Marder, {Seth R.} and Neifeld, {Mark A} and Peyghambarian, {Nasser N}",
year = "2018",
month = "12",
day = "1",
doi = "10.1038/s41467-018-04578-x",
language = "English (US)",
volume = "9",
journal = "Nature Communications",
issn = "2041-1723",
publisher = "Nature Publishing Group",
number = "1",

}

TY - JOUR

T1 - Nonlinear optical components for all-optical probabilistic graphical model

AU - Babaeian, Masoud

AU - Blanche, Pierre Alexandre

AU - Norwood, Robert A

AU - Kaplas, Tommi

AU - Keiffer, Patrick

AU - Svirko, Yuri

AU - Allen, Taylor G.

AU - Chen, Vincent W.

AU - Chi, San Hui

AU - Perry, Joseph W.

AU - Marder, Seth R.

AU - Neifeld, Mark A

AU - Peyghambarian, Nasser N

PY - 2018/12/1

Y1 - 2018/12/1

N2 - The probabilistic graphical models (PGMs) are tools that are used to compute probability distributions over large and complex interacting variables. They have applications in social networks, speech recognition, artificial intelligence, machine learning, and many more areas. Here, we present an all-optical implementation of a PGM through the sum-product message passing algorithm (SPMPA) governed by a wavelength multiplexing architecture. As a proof-of-concept, we demonstrate the use of optics to solve a two node graphical model governed by SPMPA and successfully map the message passing algorithm onto photonics operations. The essential mathematical functions required for this algorithm, including multiplication and division, are implemented using nonlinear optics in thin film materials. The multiplication and division are demonstrated through a logarithm-summation-exponentiation operation and a pump-probe saturation process, respectively. The fundamental bottlenecks for the scalability of the presented scheme are discussed as well.

AB - The probabilistic graphical models (PGMs) are tools that are used to compute probability distributions over large and complex interacting variables. They have applications in social networks, speech recognition, artificial intelligence, machine learning, and many more areas. Here, we present an all-optical implementation of a PGM through the sum-product message passing algorithm (SPMPA) governed by a wavelength multiplexing architecture. As a proof-of-concept, we demonstrate the use of optics to solve a two node graphical model governed by SPMPA and successfully map the message passing algorithm onto photonics operations. The essential mathematical functions required for this algorithm, including multiplication and division, are implemented using nonlinear optics in thin film materials. The multiplication and division are demonstrated through a logarithm-summation-exponentiation operation and a pump-probe saturation process, respectively. The fundamental bottlenecks for the scalability of the presented scheme are discussed as well.

UR - http://www.scopus.com/inward/record.url?scp=85047871767&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85047871767&partnerID=8YFLogxK

U2 - 10.1038/s41467-018-04578-x

DO - 10.1038/s41467-018-04578-x

M3 - Article

C2 - 29844343

AN - SCOPUS:85047871767

VL - 9

JO - Nature Communications

JF - Nature Communications

SN - 2041-1723

IS - 1

M1 - 2128

ER -