Photonics for Neuromorphic Computing: Fundamentals, Devices, and Opportunities

被引:5
作者
Li, Renjie [1 ]
Gong, Yuanhao [1 ]
Huang, Hai [1 ]
Zhou, Yuze [1 ]
Mao, Sixuan [1 ]
Wei, Zhijian [2 ]
Zhang, Zhaoyu [1 ]
机构
[1] Chinese Univ Hong Kong, Sch Sci & Engn, Guangdong Key Lab Optoelect Mat & Chips, Shenzhen Key Lab Semicond Lasers, Shenzhen 518172, Guangdong, Peoples R China
[2] SONT Technol Co LTD, Shenzhen 510245, Guangdong, Peoples R China
基金
中国国家自然科学基金;
关键词
artificial intelligence; large language models; neuromorphic computing; opto-electronics; photonics; photonic integrated circuits; OPTICAL NEURAL-NETWORK; ON-CHIP; MOORES LAW; 3D; NANOPHOTONICS; PERFORMANCE; SYSTEM; LASER; POWER; END;
D O I
10.1002/adma.202312825
中图分类号
O6 [化学];
学科分类号
0703 ;
摘要
In the dynamic landscape of Artificial Intelligence (AI), two notable phenomena are becoming predominant: the exponential growth of large AI model sizes and the explosion of massive amount of data. Meanwhile, scientific research such as quantum computing and protein synthesis increasingly demand higher computing capacities. As the Moore's Law approaches its terminus, there is an urgent need for alternative computing paradigms that satisfy this growing computing demand and break through the barrier of the von Neumann model. Neuromorphic computing, inspired by the mechanism and functionality of human brains, uses physical artificial neurons to do computations and is drawing widespread attention. This review studies the expansion of optoelectronic devices on photonic integration platforms that has led to significant growth in photonic computing, where photonic integrated circuits (PICs) have enabled ultrafast artificial neural networks (ANN) with sub-nanosecond latencies, low heat dissipation, and high parallelism. In particular, various technologies and devices employed in neuromorphic photonic AI accelerators, spanning from traditional optics to PCSEL lasers are examined. Lastly, it is recognized that existing neuromorphic technologies encounter obstacles in meeting the peta-level computing speed and energy efficiency threshold, and potential approaches in new devices, fabrication, materials, and integration to drive innovation are also explored. As the current challenges and barriers in cost, scalability, footprint, and computing capacity are resolved one-by-one, photonic neuromorphic systems are bound to co-exist with, if not replace, conventional electronic computers and transform the landscape of AI and scientific computing in the foreseeable future. This review examines advancements in integrated photonic neuromorphic systems, focusing on materials and device engineering breakthroughs needed to advance the field. We analyze various technologies in neuromorphic photonic AI accelerators, evaluating energy efficiency and compute density. Highlighting components like PCSEL lasers and optical interconnects, we discuss recent breakthroughs and recognize obstacles to achieving peta-level performance. Potential innovations in devices, materials, and integration are explored to overcome these challenges and transform AI and scientific computing in the near future.image
引用
收藏
页数:31
相关论文
共 193 条
  • [1] Digital in-memory stochastic computing architecture for vector-matrix multiplication
    Agwa, Shady
    Prodromakis, Themis
    [J]. FRONTIERS IN NANOTECHNOLOGY, 2023, 5
  • [2] Review of nanophotonics approaches using nanostructures and nanofabrication for III-nitrides ultraviolet-photonic devices
    Alias, Mohd Sharizal
    Tangi, Malleswararao
    Holguin-Lerma, Jorge A.
    Stegenburgs, Edgars
    Alatawi, Abdullah A.
    Ashry, Islam
    Subedi, Ram Chandra
    Priante, Davide
    Shakfa, Mohammad Khaled
    Ng, Tien Khee
    Ooi, Boon S.
    [J]. JOURNAL OF NANOPHOTONICS, 2018, 12 (04)
  • [3] AlQadasi M., 2022, IEEE J SEL TOP QUANT, V29, P2
  • [4] Advances and applications of nanophotonic biosensors
    Altug, Hatice
    Oh, Sang-Hyun
    Maier, Stefan A.
    Homola, Jiri
    [J]. NATURE NANOTECHNOLOGY, 2022, 17 (01) : 5 - 16
  • [5] 0.52 V mm ITO-based Mach-Zehnder modulator in silicon photonics
    Amin, Rubab
    Maiti, Rishi
    Carfano, Caitlin
    Ma, Zhizhen
    Tahersima, Mohammad H.
    Lilach, Yigal
    Ratnayake, Dilan
    Dalir, Hamed
    Sorger, Volker J.
    [J]. APL PHOTONICS, 2018, 3 (12)
  • [6] Arikpo I. I., 2008, Global J. Math. Sci., V6, P97
  • [7] Quantifying the Robustness of Topological Slow Light
    Arregui, Guillermo
    Gomis-Bresco, Jordi
    Sotomayor-Torres, Clivia M.
    David Garcia, Pedro
    [J]. PHYSICAL REVIEW LETTERS, 2021, 126 (02)
  • [8] An on-chip photonic deep neural network for image classification
    Ashtiani, Farshid
    Geers, Alexander J.
    Aflatouni, Firooz
    [J]. NATURE, 2022, 606 (7914) : 501 - +
  • [9] CAN PROGRAMMING BE LIBERATED FROM VON NEUMANN STYLE - FUNCTIONAL STYLE AND ITS ALGEBRA OF PROGRAMS
    BACKUS, J
    [J]. COMMUNICATIONS OF THE ACM, 1978, 21 (08) : 613 - 641
  • [10] Nonreciprocal lasing in topological cavities of arbitrary geometries
    Bahari, Babak
    Ndao, Abdoulaye
    Vallini, Felipe
    El Amili, Abdelkrim
    Fainman, Yeshaiahu
    Kante, Boubacar
    [J]. SCIENCE, 2017, 358 (6363) : 636 - 639