Introduction
For decades, information and communication technologies have exploited the properties of light to transport an ever-increasing amount of data to provide massive connectivity and broadband services. Optical fibers have replaced copper cables in long-haul transmission links, inter- and intra-datacenter communications, and fiber-to-the-home access networks. Although photonics is well-established for high-capacity transmission, its use in computing is still largely unexploited.
This is because Moore’s scaling of digital electronic processors and the use of parallel and distributed computing architectures have satisfied the need for increasingly better computer performance. However, with the advent of the deep learning era, after around 2010, computing power started doubling approximately every 4 months, which is much faster than Moore’s law ; (see Figure 1). Supercomputers are rapidly approaching the exascale era, in which requirements regarding latency, bandwidth, and energy consumption are challenging for digital electronics. In addition to this, the explosive growth of artificial intelligence (AI) and machine learning, and their penetration into everyday life, is forcing us to reconsider the traditional way computers work , and post-Moore paradigms and computer architectures need to be seriously considered . Centralized architectures are indeed inefficient in implementing models used for artificial neural networks (ANN), which are inherently distributed and require massive parallel interconnections between a multitude of elementary computing units .
FIGURE 1
Furthermore, we should consider that, even though we presume that all information is digital, “real” data, such as images, sounds, and objects, are inherently analog in their nature. If, on the one hand, digital computing does allow very complex processing, on the other hand, as the volume of data increases, digital operations become increasingly less sustainable. Today, for applications requiring massive computing, there is great interest in reconsidering analog technologies, which use dedicated circuits to efficiently process large amounts of data at very high speed .
In this scenario, photonics is emerging as a promising technology to enable the sustainable development of high-performance computing (HPC) . Not only can photonics provide fast and high-capacity interconnections among distributed processors , it can also perform calculations efficiently inside them . Mapping mathematical operations onto photonic hardware enables the implementation of accelerators for neuromorphic computing , analog optical processors ; , and quantum integrated processors ; . Several examples of programmable photonic circuits for advanced optical computing and quantum processing have been recently proposed . Multiplications and accumulations (MACs) and matrix-vector multiplications (MVMs) , which are the most time- and energy-consuming operations in ANNs, can be efficiently performed directly in the optical domain by exploiting diffractive free-space devices (Figure 2A) ; and multi-plane light converters (MPLCs), or coherent feed-forward meshes of interferometers (Figure 2B) ; ; or non-coherent filter banks based on wavelength division multiplexing (Figure 2C) ; . Such accelerators are expected to enable processing speeds of petaMAC operations per second per mm2 and energy efficiencies of attojoule per MAC ; . Recursive programmable photonic circuits with optical feedback structures have been proposed as mathematical co-processors and reconfigurable solvers of several classes of integral/differential equations (Figure 2D) .
FIGURE 2
To make these proof-of-concept photonic devices evolve toward mature large-scale computing systems, research in “neuromorphic photonics and photonic computing” needs to tackle several open challenges that vertically span from the investigation of novel material properties up to the optimization of computing networks:
• Energy sustainability. To dramatically reduce power dissipation in photonic computers that can be programmed and trained with “quasi-zero-energy-consumption”, technology breakthroughs are required to implement fast and energy-efficient optical actuators. Recently, several advances have been demonstrated by exploiting fast electro-optic and plasmonic weighting banks in photonic ANNs as well as non-volatile switchable elements based on phase-change materials Figure 3A; ; , optomechanical devices , optical memristors ; , and liquid crystal devices .
• More functionalities. Novel devices are required to enlarge the portfolio of signal processing operations that can be operated directly in the optical domain, both in free-space optics and guided wave devices. Here, promising device concepts are emerging, which are enabled by the realization of tuneable meta-surfaces, subwavelength meta-structures, and meta-devices engineered at the nanoscale Figure 3B; ; .
• Scalability. Today, photonic integrated circuits (PICs) can reach very high-integration densities enabled by complementary-metal-oxide-semiconductor (CMOS) manufacturing, but scalability to complex on-chip architectures requires sophisticated modeling and design techniques as well as advanced algorithms, control techniques, and training tools for automated calibration and adaptive reconfiguration Figure 3C; ; ; .
• Non-linear processing and data storage. Such operations, which are essential in neuromorphic computing and AI, are still inefficient when operated in optics with respect to the performance offered by electronics; in this view, the integration of photonics and electronic circuits may offer a viable route to profit from the main advantages offered by the two technological platforms. However, this approach requires reconsideration of the way integrated circuits are realized, developing suitable procedures for co-designing and co-packaging photonics and electronics chips .
• Electronic-photonic co-integration. The integration of analog photonic elements into analog/digital electronic circuits could herald new classes of computing devices combining the best of both worlds. However, such interplay opens new issues related to the co-design, co-integration, and co-packaging of photonic and electronic circuits. Monolithic integration Figure 3D; , heterogeneous integration, and chip-to-chip bonding ; are different approaches that are being explored to leverage the potential of photonic computing.
FIGURE 3
Applications for photonic computing are envisioned in mathematical accelerators for deep learning ; , photonic reservoir computing , and spiking neural networks ; cognitive radio and compressive sensing ; the classification of unstructured signals, high-energy particle collision , and qubit systems ; the non-linear control and optimization in robotics and autonomous vehicles ; intelligent signal processing in wireless and fibre-optic communication ; ; and optical quantum key distribution and chaotic systems and encryption .
Although photonic computing is still in its infancy, in recent years, we have assisted an impressive increase in fundamental and applied research in this field, which is driven by large industrial investment and the birth of small- and medium-scale enterprises. This means that photonic computing is not just a fascinating scientific area but its potential is critically strategic for stakeholders to drive investments in those areas that might have the highest impact, such as AI and quantum information. Photonic computing will probably never replace conventional computers but it can offer real opportunities for a sustainable scalability of performance in distributed and edge-computing systems. Indeed, without photonics, we may have no other alternative technologies to address the bottleneck that electronics are bound to face in meeting the requirements of future computing. .
Statements
Author contributions
FM: Writing–original draft.
Funding
The author(s) declare financial support was received for the research, authorship, and/or publication of this article. FM acknowledges financial support from the Italian National Recovery and Resilience Plan (NRRP) of NextGenerationEU, partnership on ‘Telecommunications of the Future’ (PE00000001—program ‘RESTART’, Structural Project ‘Rigoletto’ and Focused Project ‘HePIC’).
Conflict of interest
The author declares that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
The author(s) declared that they were an editorial board member of Frontiers, at the time of submission. This had no impact on the peer review process and the final decision.
Publisher’s note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
References
1
AmbrogioS.NarayananP.OkazakiA.FasoliA.MackinC.HosokawaK.et al (2023). An analog-ai chip for energy-efficient speech recognition and transcription. Nature620, 768–775. 10.1038/s41586-023-06337-5
2
AtabakiA. H.MoazeniS.PavanelloF.GevorgyanH.NotarosJ.AlloattiL.et al (2018). Integrating photonics with silicon nanoelectronics for the next generation of systems on a chip. Nature556, 349–354. 10.1038/s41586-018-0028-z
3
BogaertsW.PérezD.CapmanyJ.MillerD. A. B.PoonJ.EnglundD.et al (2020). Programmable photonic circuits. Nature586, 207–216. 10.1038/s41586-020-2764-0
4
ChangP.-H.SamantaA.YanP.FuM.ZhangY.OnM. B.et al (2023). A 3d integrated energy-efficient transceiver realized by direct bond interconnect of co-designed 12 nm finfet and silicon photonic integrated circuits. J. Light. Technol.41, 6741–6755. 10.1109/jlt.2023.3291704
5
CorrielliG.CrespiA.OsellameR. (2021). Femtosecond laser micromachining for integrated quantum photonics. Nanophotonics10, 3789–3812. 10.1515/nanoph-2021-0419
6
CuiT. J.QiM. Q.WanX.ZhaoJ.ChengQ. (2014). Coding metamaterials, digital metamaterials and programmable metamaterials. Light Sci. Appl.3, e218. 10.1038/lsa.2014.99
7
DabosG.BellasD. V.StabileR.Moralis-PegiosM.GiamougiannisG.TsakyridisA.et al (2022). Neuromorphic photonic technologies and architectures: scaling opportunities and performance frontiers [Invited]. Opt. Mat. Express12, 2343–2367. 10.1364/OME.452138
8
de LimaT. F.PengH.-T.TaitA. N.NahmiasM. A.MillerH. B.ShastriB. J.et al (2019). Machine learning with neuromorphic photonics. J. Light. Technol.37, 1515–1534. 10.1109/JLT.2019.2903474
9
DuarteJ.HanS.HarrisP.JindarianiS.KreinarE.KreisB.et al (2018). Fast inference of deep neural networks in fpgas for particle physics. J. Instrum.13, P07027. 10.1088/1748-0221/13/07/P07027
10
FarmakidisN.YoungbloodN.LeeJ. S.FeldmannJ.LodiA.LiX.et al (2022). Electronically reconfigurable photonic switches incorporating plasmonic structures and phase change materials. Adv. Sci.9, 2200383. 10.1002/advs.202200383
11
FeldmannJ.YoungbloodN.WrightC. D.BhaskaranH.PerniceW. H. P. (2019). All-optical spiking neurosynaptic networks with self-learning capabilities. Nature569, 208–214. 10.1038/s41586-019-1157-8
12
FilipovichM. J.GuoZ.Al-QadasiM.MarquezB. A.MorisonH. D.SorgerV. J.et al (2022). Silicon photonic architecture for training deep neural networks with direct feedback alignment. Optica9, 1323–1332. 10.1364/OPTICA.475493
13
GiamougiannisG.TsakyridisA.Moralis-PegiosM.PappasC.KirtasM.PassalisN.et al (2023). Analog nanophotonic computing going practical: silicon photonic deep learning engines for tiled optical matrix multiplication with dynamic precision. Nanophotonics12, 963–973. 10.1515/nanoph-2022-0423
14
HalirR.BockP. J.ChebenP.Ortega-MoñuxA.Alonso-RamosC.SchmidJ. H.et al (2015). Waveguide sub-wavelength structures: a review of principles and applications. Laser and Photonics Rev.9, 25–49. 10.1002/lpor.201400083
15
HarrisN. C.SteinbrecherG. R.PrabhuM.LahiniY.MowerJ.BunandarD.et al (2017). Quantum transport simulations in a programmable nanophotonic processor. Nat. Photonics11, 447–452. 10.1038/nphoton.2017.95
16
HuangC.SorgerV. J.MiscuglioM.Al-QadasiM.MukherjeeA.LampeL.et al (2022). Prospects and applications of photonic neural networks. Adv. Phys. X7, 1981155. 10.1080/23746149.2021.1981155
17
HughesT. W.MinkovM.ShiY.FanS. (2018). Training of photonic neural networks through in situ backpropagation and gradient measurement. Optica5, 864–871. 10.1364/OPTICA.5.000864
18
JainA.MaoJ.MohiuddinK. (1996). Artificial neural networks: a tutorial. Computer29, 31–44. 10.1109/2.485891
19
KilicV.TranT. D.FosterM. A. (2023). Compressed sensing in photonics: tutorial. J. Opt. Soc. Am. B40, 28–52. 10.1364/JOSAB.469865
20
KitayamaK.-i.NotomiM.NaruseM.InoueK.KawakamiS.UchidaA. (2019). Novel frontier of photonics for data processing—photonic accelerator. Apl. Photonics4, 090901. 10.1063/1.5108912
21
KuuttiS.BowdenR.JinY.BarberP.FallahS. (2021). A survey of deep learning applications to autonomous vehicle control. IEEE Trans. Intelligent Transp. Syst.22, 712–733. 10.1109/TITS.2019.2962338
22
KwekL.-C.CaoL.LuoW.WangY.SunS.WangX.et al (2021). Chip-based quantum key distribution. AAPPS Bull.31, 15–4710. 10.1007/s43673-021-00017-0
23
LiL.ZhaoH.LiuC.LiL.CuiT. J. (2022). Intelligent metasurfaces: control, communication and computing. eLight2, 7–8643. 10.1186/s43593-022-00013-3
24
LinX.RivensonY.YardimciN. T.VeliM.LuoY.JarrahiM.et al (2018). All-optical machine learning using diffractive deep neural networks. Science361, 1004–1008. 10.1126/science.aat8084
25
MagesanE.GambettaJ. M.CórcolesA. D.ChowJ. M. (2015). Machine learning for discriminating quantum measurement trajectories and improving readout. Phys. Rev. Lett.114, 200501. 10.1103/PhysRevLett.114.200501
26
MenguD.ZhaoY.YardimciN. T.RivensonY.JarrahiM.OzcanA. (2020). Misalignment resilient diffractive optical networks. Nanophotonics9, 4207–4219. 10.1515/nanoph-2020-0291
27
MilanizadehM.AhmadiS.PetriniM.AguiarD.MazzantiR.ZanettoF.et al (2020). Control and calibration recipes for photonic integrated circuits. IEEE J. Sel. Top. Quantum Electron.26, 1–10. 10.1109/JSTQE.2020.2975657
28
MilanizadehM.SeyedinNavadehS.ZanettoF.GrimaldiV.VitaC. D.KlitisC.et al (2022). Separating arbitrary free-space beams with an integrated photonic processor. Light Sci. Appl.11, 197. 10.1038/s41377-022-00884-8
29
MillerD. A. B. (2017). Attojoule optoelectronics for low-energy information processing and communications. J. Light. Technol.35, 346–396. 10.1109/JLT.2017.2647779
30
MiscuglioM.SorgerV. J. (2020). Photonic tensor cores for machine learning. Appl. Phys. Rev.7, 031404. 10.1063/5.0001942
31
NahmiasM. A.de LimaT. F.TaitA. N.PengH.-T.ShastriB. J.PrucnalP. R. (2020). Photonic multiply-accumulate operations for neural networks. IEEE J. Sel. Top. Quantum Electron.26, 1–18. 10.1109/JSTQE.2019.2941485
32
NakajimaM.TanakaK.HashimotoT. (2021). Scalable reservoir computing on coherent linear photonic processor. Commun. Phys.4, 20–3650. 10.1038/s42005-021-00519-1
33
NezamiM. S.de LimaT. F.MitchellM.YuS.WangJ.BilodeauS.et al (2023). Packaging and interconnect considerations in neuromorphic photonic accelerators. IEEE J. Sel. Top. Quantum Electron.29, 1–11. 10.1109/JSTQE.2022.3200604
34
NozakiK.MatsuoS.FujiiT.TakedaK.ShinyaA.KuramochiE.et al (2019). Femtofarad optoelectronic integration demonstrating energy-saving signal conversion and nonlinear functions. Nat. Photonics13, 454–459. 10.1038/s41566-019-0397-3
35
PaiS.ParkT.BallM.PenkovskyB.DubrovskyM.AbebeN.et al (2023a). Experimental evaluation of digitally verifiable photonic computing for blockchain and cryptocurrency. Optica10, 552–560. 10.1364/OPTICA.476173
36
PaiS.SunZ.HughesT. W.ParkT.BartlettB.WilliamsonI. A. D.et al (2023b). Experimentally realized in situ backpropagation for deep learning in photonic neural networks. Science380, 398–404. 10.1126/science.ade8450
37
PesericoN.de LimaT. F.PrucnalP.SorgerV. J. (2022). Emerging devices and packaging strategies for electronic-photonic ai accelerators: opinion. Opt. Mat. Express12, 1347–1351. 10.1364/OME.451802
38
QuackN.TakabayashiA. Y.SattariH.EdingerP.JoG.BleikerS. J.et al (2023). Integrated silicon photonic mems. Microsystems Nanoeng.9, 27–7434. 10.1038/s41378-023-00498-z
39
SevillaJ.HeimL.HoA.BesirogluT.HobbhahnM.VillalobosP. (2022). “Compute trends across three eras of machine learning,” in 2022 International Joint Conference on Neural Networks (IJCNN), 1–8. 10.1109/IJCNN55064.2022.9891914
40
ShalfJ. (2020). The future of computing beyond Moore’s Law. Philosophical Trans. R. Soc. A Math. Phys. Eng. Sci.378, 20190061. 10.1098/rsta.2019.0061
41
ShastriB. J.TaitA. N.Ferreira de LimaT.PerniceW. H. P.BhaskaranH.WrightC. D.et al (2021). Photonics for artificial intelligence and neuromorphic computing. Nat. Photonics15, 102–114. 10.1038/s41566-020-00754-y
42
ShenY.HarrisN. C.SkirloS.PrabhuM.Baehr-JonesT.HochbergM.et al (2017). Deep learning with coherent nanophotonic circuits. Nat. Photonics11, 441–446. 10.1038/nphoton.2017.93
43
StroevN.BerloffN. G. (2023). Analog photonics computing for information processing, inference, and optimization. Adv. Quantum Technol.6, 2300055. 10.1002/qute.202300055
44
TaitA. N.JayatillekaH.LimaT. F. D.MaP. Y.NahmiasM. A.ShastriB. J.et al (2018). Feedback control for microring weight banks. Opt. Express26, 26422–26443. 10.1364/OE.26.026422
45
ThraskiasC. A.LallasE. N.NeumannN.ScharesL.OffreinB. J.HenkerR.et al (2018). Survey of photonic and plasmonic interconnect technologies for intra-datacenter and high-performance computing communications. IEEE Commun. Surv. Tutorials20, 2758–2783. 10.1109/COMST.2018.2839672
46
TzarouchisD. C.MencagliM. J.EdwardsB.EnghetaN. (2022). Mathematical operations and equation solving with reconfigurable metadevices. Light Sci. Appl.11, 263. 10.1038/s41377-022-00950-1
47
WuJ.LinX.GuoY.LiuJ.FangL.JiaoS.et al (2022). Analog optical computing for artificial intelligence. Engineering10, 133–145. 10.1016/j.eng.2021.06.021
48
XuX.-Y.JinX.-M. (2023). Integrated photonic computing beyond the von neumann architecture. ACS Photonics10, 1027–1036. 10.1021/acsphotonics.2c01543
49
YinK.HsiangE.-L.ZouJ.LiY.YangZ.YangQ.et al (2022). Advanced liquid crystal devices for augmented reality and virtual reality displays: principles and applications. Light Sci. Appl.11, 161. 10.1038/s41377-022-00851-3
50
YoungbloodN.Ríos OcampoC. A.PerniceW. H. P.BhaskaranH. (2023). Integrated optical memristors. Nat. Photonics17, 561–572. 10.1038/s41566-023-01217-w
51
ZanettoF.TosoF.GrimaldiV.PetriniM.MartinezA.MilanizadehM.et al (2023). Time-multiplexed control of programmable silicon photonic circuits enabled by monolithic cmos electronics. Laser and Photonics Rev.17, 2300124. 10.1002/lpor.202300124
52
ZhangH.GuM.JiangX. D.ThompsonJ.CaiH.PaesaniS.et al (2021). An optical neural chip for implementing complex-valued neural network. Nat. Commun.12, 457. 10.1038/s41467-020-20719-7
53
ZhangY.NaumanU. (2020). Deep learning trends driven by temes: a philosophical perspective. IEEE Access8, 196587–196599. 10.1109/ACCESS.2020.3032143
54
ZhouH.DongJ.ChengJ.DongW.HuangC.ShenY.et al (2022). Photonic matrix multiplication lights up photonic accelerator and beyond. Light Sci. Appl.11, 30. 10.1038/s41377-022-00717-8
55
ZhuD.PanS. (2020). Broadband cognitive radio enabled by photonics. J. Light. Technol.38, 3076–3088. 10.1109/jlt.2020.2993021
Summary
Keywords
photonic computing, photonic integrated circuits (PICs), photonic neural networks, photonic quantum computing, analog computing, photonics, neuromorphic photonics
Citation
Morichetti F (2024) Grand challenges in neuromorphic photonics and photonic computing. Front. Photonics 4:1336510. doi: 10.3389/fphot.2023.1336510
Received
10 November 2023
Accepted
30 December 2023
Published
29 January 2024
Volume
4 - 2023
Edited and reviewed by
Marco Peccianti, Loughborough University, United Kingdom
Updates
Copyright
© 2024 Morichetti.
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
*Correspondence: Francesco Morichetti, francesco.morichetti@polimi.it
Disclaimer
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.