Photonic Neural Networks: A Survey DOI Creative Commons
Lorenzo De Marinis, Marco Cococcioni, P. Castoldi

et al.

IEEE Access, Journal Year: 2019, Volume and Issue: 7, P. 175827 - 175841

Published: Jan. 1, 2019

Photonic solutions are today a mature industrial reality concerning high speed, throughput data communication and switching infrastructures. It is still matter of investigation to what extent photonics will play role in next-generation computing architectures. In particular, due the recent outstanding achievements artificial neural networks, there big interest trying improve their speed energy efficiency by exploiting photonic-based hardware instead electronic-based hardware. this work we review state-of-the-art photonic networks. We propose taxonomy existing (categorized into multilayer perceptrons, convolutional spiking reservoir computing) with emphasis on proof-of-concept implementations. also survey specific approaches developed for training Finally discuss open challenges highlight most promising future research directions field.

Language: Английский

Photonics for artificial intelligence and neuromorphic computing DOI
Bhavin J. Shastri, Alexander N. Tait, Thomas Ferreira de Lima

et al.

Nature Photonics, Journal Year: 2021, Volume and Issue: 15(2), P. 102 - 114

Published: Jan. 29, 2021

Language: Английский

Citations

1253
Programmable photonic circuits DOI
Wim Bogaerts, Daniel Pérez, J. Capmany

et al.

Nature, Journal Year: 2020, Volume and Issue: 586(7828), P. 207 - 216

Published: Oct. 7, 2020

Language: Английский

Citations

917
An optical neural chip for implementing complex-valued neural network DOI Creative Commons
Hui Zhang, Mile Gu, Xudong Jiang

et al.

Nature Communications, Journal Year: 2021, Volume and Issue: 12(1)

Published: Jan. 19, 2021

Abstract Complex-valued neural networks have many advantages over their real-valued counterparts. Conventional digital electronic computing platforms are incapable of executing truly complex-valued representations and operations. In contrast, optical that encode information in both phase magnitude can execute complex arithmetic by interference, offering significantly enhanced computational speed energy efficiency. However, to date, most demonstrations still only utilize conventional frameworks designed for computers, forfeiting the such as efficient this article, we highlight an chip (ONC) implements networks. We benchmark performance our ONC four settings: simple Boolean tasks, species classification Iris dataset, classifying nonlinear datasets (Circle Spiral), handwriting recognition. Strong learning capabilities (i.e., high accuracy, fast convergence capability construct decision boundaries) achieved compared its counterpart.

Language: Английский

Citations

462
A New Family of Ultralow Loss Reversible Phase‐Change Materials for Photonic Integrated Circuits: Sb2S3 and Sb2Se3 DOI Creative Commons
Matthew Delaney, Ioannis Zeimpekis, Daniel Lawson

et al.

Advanced Functional Materials, Journal Year: 2020, Volume and Issue: 30(36)

Published: July 9, 2020

Abstract Phase‐change materials (PCMs) are seeing tremendous interest for their use in reconfigurable photonic devices; however, the most common PCMs exhibit a large absorption loss one or both states. Here, Sb 2 S 3 and Se demonstrated as class of low loss, reversible alternatives to standard commercially available chalcogenide PCMs. A contrast refractive index Δ n = 0.60 0.77 is reported, while maintaining very losses ( k < 10 −5 ) telecommunications C‐band at 1550 nm. With stronger visible spectrum, allows optical switching using conventional wavelength lasers. stable endurance better than 4000 cycles demonstrated. To deal with essentially zero intrinsic losses, new figure merit (FOM) introduced taking into account measured waveguide when integrating these onto silicon photonics platform. The FOM 29 rad phase shift per dB outperforms Ge Te 5 by two orders magnitude paves way on‐chip programmable control. These truly low‐loss switchable open up directions integrated circuits, metasurfaces, nanophotonic devices.

Language: Английский

Citations

430
Tunable nanophotonics enabled by chalcogenide phase-change materials DOI Creative Commons
Sajjad Abdollahramezani, Omid Hemmatyar, Hossein Taghinejad

et al.

Nanophotonics, Journal Year: 2020, Volume and Issue: 9(5), P. 1189 - 1241

Published: May 1, 2020

Nanophotonics has garnered intensive attention due to its unique capabilities in molding the flow of light subwavelength regime. Metasurfaces (MSs) and photonic integrated circuits (PICs) enable realization mass-producible, cost-effective, highly efficient flat optical components for imaging, sensing, communications. In order nanophotonics with multi-purpose functionalities, chalcogenide phase-change materials (PCMs) have been introduced as a promising platform tunable reconfigurable nanophotonic frameworks. Integration non-volatile PCMs properties such drastic contrasts, fast switching speeds, long-term stability grants substantial reconfiguration more conventional static platforms. this review, we discuss state-of-the-art developments well emerging trends MSs PICs using PCMs. We outline material properties, structural transformation, electro-optic, thermo-optic effects well-established classes The deep learning-based approaches optimization analysis light-matter interactions are also discussed. review is concluded by discussing existing challenges adjustable perspective on possible area.

Language: Английский

Citations

369
Photonic matrix multiplication lights up photonic accelerator and beyond DOI Creative Commons
Hailong Zhou, Jianji Dong, Junwei Cheng

et al.

Light Science & Applications, Journal Year: 2022, Volume and Issue: 11(1)

Published: Feb. 3, 2022

Abstract Matrix computation, as a fundamental building block of information processing in science and technology, contributes most the computational overheads modern signal artificial intelligence algorithms. Photonic accelerators are designed to accelerate specific categories computing optical domain, especially matrix multiplication, address growing demand for resources capacity. multiplication has much potential expand domain telecommunication, benefiting from its superior performance. Recent research photonic flourished may provide opportunities develop applications that unachievable at present by conventional electronic processors. In this review, we first introduce methods mainly including plane light conversion method, Mach–Zehnder interferometer method wavelength division multiplexing method. We also summarize developmental milestones related applications. Then, review their detailed advances neural networks recent years. Finally, comment on challenges perspectives acceleration.

Language: Английский

Citations

342
Large-scale integration of artificial atoms in hybrid photonic circuits DOI
Noel Wan, Tsung‐Ju Lu, Kevin C. Chen

et al.

Nature, Journal Year: 2020, Volume and Issue: 583(7815), P. 226 - 231

Published: July 8, 2020

Language: Английский

Citations

339
Integrated lithium niobate electro-optic modulators: when performance meets scalability DOI Creative Commons
Mian Zhang, Cheng Wang, Prashanta Kharel

et al.

Optica, Journal Year: 2021, Volume and Issue: 8(5), P. 652 - 652

Published: March 17, 2021

Electro-optic modulators (EOMs) convert signals from the electrical to optical domain. They are at heart of communication, microwave signal processing, sensing, and quantum technologies. Next-generation EOMs require high-density integration, low cost, high performance simultaneously, which difficult achieve with established integrated photonics platforms. Thin-film lithium niobate (LN) has recently emerged as a strong contender owing its intrinsic electro-optic (EO) efficiency, industry-proven performance, robustness, and, importantly, rapid development scalable fabrication techniques. The thin-film LN platform inherits nearly all material advantages legacy bulk devices amplifies them smaller footprint, wider bandwidths, lower power consumption. Since first adoption commercial wafers only few years ago, overall is already comparable with, if not exceeding, best alternatives based on mature platforms such silicon indium phosphide, have benefited many decades research development. In this mini-review, we explain principles technical advances that enabled state-of-the-art modulator demonstrations. We discuss several approaches, their challenges. also outline paths follow improve further, provide perspective what believe could become in future. Finally, key subcomponent more complex photonic functionalities, look forward exciting opportunities for larger-scale EO circuits beyond single components.

Language: Английский

Citations

316
Quantum networks based on color centers in diamond DOI Creative Commons
Maximilian Ruf, Noel Wan,

Hyeongrak Choi

et al.

Journal of Applied Physics, Journal Year: 2021, Volume and Issue: 130(7)

Published: Aug. 16, 2021

With the ability to transfer and process quantum information, large-scale networks will enable a suite of fundamentally new applications, from communications distributed sensing, metrology, computing. This Perspective reviews requirements for network nodes color centers in diamond as suitable node candidates. We give brief overview state-of-the-art experiments employing discuss future research directions, focusing, particular, on control coherence qubits that distribute store entangled states, efficient spin–photon interfaces. route toward integrated devices combining with other photonic materials an outlook realistic protocol implementations applications.

Language: Английский

Citations

222
Space-efficient optical computing with an integrated chip diffractive neural network DOI Creative Commons
Hanqing Zhu, Jun Zou, Hui Zhang

et al.

Nature Communications, Journal Year: 2022, Volume and Issue: 13(1)

Published: Feb. 24, 2022

Abstract Large-scale, highly integrated and low-power-consuming hardware is becoming progressively more important for realizing optical neural networks (ONNs) capable of advanced computing. Traditional experimental implementations need N 2 units such as Mach-Zehnder interferometers (MZIs) an input dimension to realize typical computing operations (convolutions matrix multiplication), resulting in limited scalability consuming excessive power. Here, we propose the diffractive network implementing parallel Fourier transforms, convolution application-specific using two ultracompact cells (Fourier transform operation) only MZIs. The footprint energy consumption scales linearly with data dimension, instead quadratic scaling traditional ONN framework. A ~10-fold reduction both consumption, well equal high accuracy previous MZI-based ONNs was experimentally achieved computations performed on MNIST Fashion-MNIST datasets. (IDNN) chip demonstrates a promising avenue towards scalable low-power-consumption computational chips optical-artificial-intelligence.

Language: Английский

Citations

215