The Review of Laser Engineering
Online ISSN : 1349-6603
Print ISSN : 0387-0200
ISSN-L : 0387-0200
Volume 51, Issue 10
Special Issue on Recent Trends in Optical Computing
Displaying 1-7 of 7 articles from this issue
Special Issue on Recent Trends in Optical Computing
Special Issue
Laser Review
  • Kenichi HIROSAWA
    2023Volume 51Issue 10 Pages 612-
    Published: 2023
    Released on J-STAGE: August 18, 2025
    JOURNAL FREE ACCESS
    Although optical computing is a research field with a long history, it has attracted much attention again in recent years. The computational demands from society are dramatically increasing due to the rapid progress of Artificial Intelligence (AI) technology, expansion of communication capacity, and utilization of big data, leading to the need for drastic improvement of power efficiency of computation. In response to these social demands, optical computing has the advantage of being able to perform certain types of operations with extremely high efficiency and/or with very low power consumption. This special issue covers the latest trends in optical computing with specific examples such as optical accelerators, quantum Ising machines, and quantum computers.
    Download PDF (356K)
  • Akihiko SHINYA
    2023Volume 51Issue 10 Pages 614-
    Published: 2023
    Released on J-STAGE: August 18, 2025
    JOURNAL FREE ACCESS
    We review opto-electronic integration technologies for creating a new computing platform with a wide bandwidth and low energy consumption. The development of large-scale silicon photonics circuits and ultra-low energy nanophotonic devices will allow for diverse types of photonic-information processing. This article introduces the technologies for high bandwidth density optical interconnection toward 1 Byte-per-FLOP and presents a path to opto-electronic integrated accelerators that will exceed current digital electronic systems by >10 2 in compute density and energy efficiency.
    Download PDF (747K)
  • Takuo TANEMURA, Ryota TANOMURA, Rui TANG, Yoshiaki NAKANO
    2023Volume 51Issue 10 Pages 619-
    Published: 2023
    Released on J-STAGE: August 18, 2025
    JOURNAL FREE ACCESS
    This paper reviews recent progress toward developing universal optical unitary converting (OUC) circuits based on the concept of multi-plane light conversion (MPLC). Unlike conventional architectures using 2 × 2 Mach-Zehnder interferometers, an MPLC-based OUC consists of multiport mixing layers that provide strong all-to-all coupling among every mode. Such inherent redundancy of MPLC provides unique scalability and excellent robustness against fabrication imperfectness, enabling the integration of large-scale OUCs on silicon and indium phosphide platforms. Moreover, we numerically demonstrate that deep neural networks employing MPLC-based OUCs show excellent performance with 95% data- classification accuracy even when the number of phase-shifting stages is reduced by 1/10. Such a property cannot be obtained by conventional schemes, revealing the unique advantage of using MPLCbased OUCs to realize large-scale optical linear transformations on a compact integrated chip.
    Download PDF (1083K)
  • Mitsuru TAKENAKA
    2023Volume 51Issue 10 Pages 624-
    Published: 2023
    Released on J-STAGE: August 18, 2025
    JOURNAL FREE ACCESS
    Heterogeneous integration is discussed to realize a large-scale, energy-efficient programmable Si photonic circuit for deep learning. A low-loss, compact, and non-volatile optical phase shifter, that is indispensable for reconfiguring a photonic circuit, is achieved using a wide-gap phase change material, GeSbTeS. An InGaAs membrane integrated on a Si optical waveguide enables ultrahigh-reponsitivity waveguide- coupled phototransistor which can be used for monitoring an optical power in a waveguide. An InGaAs membrane on a Si slot optical waveguide also makes it possible to achieve high responsivity and low capacitance simultaneously, contributing to realizing a low-power optical readout through the receiver- less system.
    Download PDF (2035K)
  • Guangwei CONG, Koji YAMADA
    2023Volume 51Issue 10 Pages 629-
    Published: 2023
    Released on J-STAGE: August 18, 2025
    JOURNAL FREE ACCESS
    Leveraging integrated photonics to implement AI and computing tasks has been attracting wide attention due to the ever-growing demand on low-latency and low-power computing resources for AI and machine learning workloads in data centers. This offers unprecedented opportunities, as well as challenges, for currently accessible photonic technologies. Different from high-speed transceivers for optical interconnects where controversy remains for selecting different technology solutions, it reaches extensive consensus that silicon photonic process technology will be the most promising solution for optical computing from several aspects such as large scale, integration with electronics, reconfigurability, interoperability, and miniaturization. In this review, we first introduce our work on utilizing silicon photonic circuits to perform novel classification computing and its recent extension to electro-photonic dynamic system, and then discuss our perspectives on the importance, robustness, and challenges of silicon photonic process technology to implement optical computing.
    Download PDF (729K)
  • Yasuhiro YAMADA, Takahiro INAGAKI
    2023Volume 51Issue 10 Pages 634-
    Published: 2023
    Released on J-STAGE: August 18, 2025
    JOURNAL FREE ACCESS
    Recently, various physical devices have been developed to solve specific problems which require large computing resources on digital computers. In this paper, we briefly introduce the development history of physical solvers for combinatorial optimization and explain Coherent Ising machines (CIMs) based on optical oscillator networks. For large optimization problems, CIMs have experimentally shown better performance than a numerical algorithm on digital computers. Furthermore, CIM applications are expanding from combinatorial optimizations to simulations and samplings. CIMs can deliver canonical and power-law distributions of Ising spins, providing a successive simulation of magnetic phase transition and potential applications for samplings in machine learning algorithms. Spiking neural networks are also simulated with a similar platform, where synchronized neurons show a spontaneous change between two firing modes observed in brain neurons. The neuro-inspired platform can be useful both for combinatorial optimization and for the analysis of brain functions and brain diseases.
    Download PDF (1029K)
  • Mamoru ENDO, Akira FURUSAWA
    2023Volume 51Issue 10 Pages 639-
    Published: 2023
    Released on J-STAGE: August 18, 2025
    JOURNAL FREE ACCESS
    Although significant progress has been made in quantum computer research over the past few years, further technological innovation is essential to achieve one of our goals: a fault-tolerant quantum computer. As research is being conducted in various physical systems, one method is attracting attention: encoding quantum information in the quadrature phase amplitude of an electromagnetic field of light. By importing optical communication technology, this time-domain multiplexing method maximizes the properties of light, a traveling wave with a high carrier frequency, and can easily solve the issues of high speed and scalability, both of which are major barriers for other methods. Research is also addressing the biggest challenge of optical methods: the generation of a non-Gaussian state for providing error correction capability. In this paper, we focus on the example of non-Gaussian state generation from the overview of optical quantum computing systems.
    Download PDF (1183K)
feedback
Top