Novel 2D Materials Enabled All-Optical Nonlinear Activation Functions for On-Chip Photonic Deep Neural Networks

A. Karabchevsky

Novel 2D Materials Enabled All-Optical Nonlinear Activation Functions for On-Chip Photonic Deep Neural Networks

School of Sustainability and Climate Change

Research output: Contribution to journal › Conference article › peer-review

Abstract

Neural networks are one of the first major milestones in developing artificial intelligence systems. The utilisation of integrated photonics in neural networks offers a promising alternative approach to microelectronic and hybrid optical-electronic implementations due to improvements in computational speed and low energy consumption in machine-learning tasks. However, at present, the optical realisation of the nonlinear activation function is a tremendous challenge.

Original language	English
Pages (from-to)	270
Number of pages	1
Journal	International Conference on Metamaterials, Photonic Crystals and Plasmonics
State	Published - 1 Jan 2022
Event	12th International Conference on Metamaterials, Photonic Crystals and Plasmonics, META 2022 - Torremolinos, Spain Duration: 19 Jul 2022 → 22 Jul 2022

ASJC Scopus subject areas

Electrical and Electronic Engineering
Materials Science (miscellaneous)
Electronic, Optical and Magnetic Materials
Materials Chemistry

Cite this

@article{14d7a3c2a6a54d4a93360bebbf707c70,

title = "Novel 2D Materials Enabled All-Optical Nonlinear Activation Functions for On-Chip Photonic Deep Neural Networks",

abstract = "Neural networks are one of the first major milestones in developing artificial intelligence systems. The utilisation of integrated photonics in neural networks offers a promising alternative approach to microelectronic and hybrid optical-electronic implementations due to improvements in computational speed and low energy consumption in machine-learning tasks. However, at present, the optical realisation of the nonlinear activation function is a tremendous challenge.",

author = "A. Karabchevsky",

note = "Publisher Copyright: {\textcopyright} 2023, META Conference. All rights reserved.; 12th International Conference on Metamaterials, Photonic Crystals and Plasmonics, META 2022 ; Conference date: 19-07-2022 Through 22-07-2022",

year = "2022",

month = jan,

day = "1",

language = "English",

pages = "270",

journal = "International Conference on Metamaterials, Photonic Crystals and Plasmonics",

issn = "2429-1390",

publisher = "META Conference",

}

TY - JOUR

T1 - Novel 2D Materials Enabled All-Optical Nonlinear Activation Functions for On-Chip Photonic Deep Neural Networks

AU - Karabchevsky, A.

PY - 2022/1/1

Y1 - 2022/1/1

N2 - Neural networks are one of the first major milestones in developing artificial intelligence systems. The utilisation of integrated photonics in neural networks offers a promising alternative approach to microelectronic and hybrid optical-electronic implementations due to improvements in computational speed and low energy consumption in machine-learning tasks. However, at present, the optical realisation of the nonlinear activation function is a tremendous challenge.

AB - Neural networks are one of the first major milestones in developing artificial intelligence systems. The utilisation of integrated photonics in neural networks offers a promising alternative approach to microelectronic and hybrid optical-electronic implementations due to improvements in computational speed and low energy consumption in machine-learning tasks. However, at present, the optical realisation of the nonlinear activation function is a tremendous challenge.

UR - http://www.scopus.com/inward/record.url?scp=85179875560&partnerID=8YFLogxK

M3 - Conference article

AN - SCOPUS:85179875560

SN - 2429-1390

SP - 270

JO - International Conference on Metamaterials, Photonic Crystals and Plasmonics

JF - International Conference on Metamaterials, Photonic Crystals and Plasmonics

T2 - 12th International Conference on Metamaterials, Photonic Crystals and Plasmonics, META 2022

Y2 - 19 July 2022 through 22 July 2022

ER -

Novel 2D Materials Enabled All-Optical Nonlinear Activation Functions for On-Chip Photonic Deep Neural Networks

Abstract

ASJC Scopus subject areas

Other files and links

Fingerprint

Cite this