{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,5]],"date-time":"2026-05-05T18:49:19Z","timestamp":1778006959005,"version":"3.51.4"},"reference-count":56,"publisher":"IOP Publishing","issue":"2","license":[{"start":{"date-parts":[[2022,6,20]],"date-time":"2022-06-20T00:00:00Z","timestamp":1655683200000},"content-version":"vor","delay-in-days":19,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"},{"start":{"date-parts":[[2022,6,20]],"date-time":"2022-06-20T00:00:00Z","timestamp":1655683200000},"content-version":"tdm","delay-in-days":19,"URL":"https:\/\/iopscience.iop.org\/info\/page\/text-and-data-mining"}],"funder":[{"DOI":"10.13039\/501100013209","name":"Hellenic Foundation for Research and Innovation","doi-asserted-by":"crossref","award":["4233"],"award-info":[{"award-number":["4233"]}],"id":[{"id":"10.13039\/501100013209","id-type":"DOI","asserted-by":"crossref"}]}],"content-domain":{"domain":["iopscience.iop.org"],"crossmark-restriction":false},"short-container-title":["Neuromorph. Comput. Eng."],"published-print":{"date-parts":[[2022,6,1]]},"abstract":"<jats:title>Abstract<\/jats:title>\n               <jats:p>Non-von-Neumann computing architectures and deep learning training models have sparked a new computational era where neurons are forming the main architectural backbone and vector, matrix and tensor multiplications comprise the basic mathematical toolbox. This paradigm shift has triggered a new race among hardware technology candidates; within this frame, the field of neuromorphic photonics promises to convolve the targeted algebraic portfolio along a computational circuitry with unique speed, parallelization, and energy efficiency advantages. Fueled by the inherent energy efficient analog matrix multiply operations of optics, the staggering advances of photonic integration and the enhanced multiplexing degrees offered by light, neuromorphic photonics has stamped the resurgence of optical computing brining a unique perspective in low-energy and ultra-fast linear algebra functions. However, the field of neuromorphic photonics has relied so far on two basic architectural schemes, i.e., coherent linear optical circuits and incoherent WDM approaches, where wavelengths have still not been exploited as a new mathematical dimension. In this paper, we present a radically new approach for promoting the synergy of WDM with universal linear optics and demonstrate a new, high-fidelity crossbar-based neuromorphic photonic platform, able to support matmul with multidimensional operands. Going a step further, we introduce the concept of programmable input and weight banks, supporting <jats:italic>in situ<\/jats:italic> reconfigurability, forming in this way the first WDM-equipped universal linear optical operator and demonstrating different operational modes like matrix-by-matrix and vector-by-tensor multiplication. The benefits of our platform are highlighted in a fully convolutional neural network layout that is responsible for parity identification in the MNIST handwritten digit dataset, with physical layer simulations revealing an accuracy of \u223c94%, degraded by only 2% compared to respective results obtained when executed entirely by software. Finally, our in-depth analysis provides the guidelines for neuromorphic photonic processor performance improvement, revealing along the way that 4 bit quantization is sufficient for inputs, whereas the weights can be implemented with as low as 2 bits of precision, offering substantial benefits in terms of driving circuitry complexity and energy savings.<\/jats:p>","DOI":"10.1088\/2634-4386\/ac724d","type":"journal-article","created":{"date-parts":[[2022,5,23]],"date-time":"2022-05-23T22:15:37Z","timestamp":1653344137000},"page":"024010","update-policy":"https:\/\/doi.org\/10.1088\/crossmark-policy","source":"Crossref","is-referenced-by-count":30,"title":["WDM equipped universal linear optics for programmable neuromorphic photonic processors"],"prefix":"10.1088","volume":"2","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-0267-7368","authenticated-orcid":true,"given":"Angelina","family":"Totovic","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-0466-9516","authenticated-orcid":true,"given":"Christos","family":"Pappas","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8670-0248","authenticated-orcid":true,"given":"Manos","family":"Kirtas","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-1498-4904","authenticated-orcid":true,"given":"Apostolos","family":"Tsakyridis","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-6406-9723","authenticated-orcid":true,"given":"George","family":"Giamougiannis","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-1177-9139","authenticated-orcid":true,"given":"Nikolaos","family":"Passalis","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-9401-730X","authenticated-orcid":true,"given":"Miltiadis","family":"Moralis-Pegios","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-1288-3667","authenticated-orcid":true,"given":"Anastasios","family":"Tefas","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2931-4540","authenticated-orcid":true,"given":"Nikos","family":"Pleros","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"266","published-online":{"date-parts":[[2022,6,20]]},"reference":[{"key":"nceac724dbib1","doi-asserted-by":"publisher","DOI":"10.1063\/1.5129306","article-title":"The building blocks of a brain-inspired computer","volume":"7","author":"Kendall","year":"2020","journal-title":"Appl. Phys. Rev."},{"key":"nceac724dbib2","article-title":"Language models are few-shot learners","author":"Brown","year":"2020"},{"key":"nceac724dbib3","article-title":"Using DeepSpeed and megatron to train megatron-turing NLG 530B, the world\u2019s largest and most powerful generative language model","author":"Kharya","year":"2021"},{"key":"nceac724dbib4","article-title":"Benchmarking TPU, GPU, and CPU platforms for deep learning","author":"Wang","year":"2019"},{"key":"nceac724dbib5","doi-asserted-by":"publisher","first-page":"354","DOI":"10.1016\/j.patcog.2017.10.013","article-title":"Recent advances in convolutional neural networks","volume":"77","author":"Gu","year":"2018","journal-title":"Pattern Recognit."},{"key":"nceac724dbib6","first-page":"6869","article-title":"Quantized neural networks: training neural networks with low precision weights and activations","volume":"18","author":"Hubara","year":"2017","journal-title":"J. Mach. Learn. Res."},{"key":"nceac724dbib7","first-page":"1556","article-title":"Training noise-resilient recurrent photonic networks for financial time series analysis","author":"Passalis","year":"2020"},{"key":"nceac724dbib8","article-title":"Noise-resilient and high-speed deep learning with coherent silicon photonics","author":"Mourgias-Alexandris","year":"2021","journal-title":"Nat. Commun."},{"key":"nceac724dbib9","doi-asserted-by":"publisher","first-page":"3243","DOI":"10.1109\/jlt.2022.3171831","article-title":"Neuromorphic silicon photonics and hardware-aware deep learning for high-speed inference","volume":"40","author":"Moralis-Pegios","year":"2022","journal-title":"J. Lightwave Technol."},{"key":"nceac724dbib10","article-title":"Quantization-aware training for low precision photonic neural networks","author":"Kirtas","year":"2021","journal-title":"Neural Netw."},{"key":"nceac724dbib11","doi-asserted-by":"publisher","first-page":"3692","DOI":"10.3390\/app11083692","article-title":"Training and inference of optical neural networks with noise and low-bits control","volume":"11","author":"Zhang","year":"2021","journal-title":"Appl. Sci."},{"key":"nceac724dbib12"},{"key":"nceac724dbib13"},{"key":"nceac724dbib14"},{"key":"nceac724dbib15"},{"key":"nceac724dbib16","doi-asserted-by":"publisher","first-page":"1623","DOI":"10.1364\/optica.5.001623","article-title":"Linear programmable nanophotonic processors","volume":"5","author":"Harris","year":"2018","journal-title":"Optica"},{"key":"nceac724dbib17"},{"key":"nceac724dbib18","doi-asserted-by":"publisher","first-page":"441","DOI":"10.1038\/nphoton.2017.93","article-title":"Deep learning with coherent nanophotonic circuits","volume":"11","author":"Shen","year":"2017","journal-title":"Nat. Photon."},{"key":"nceac724dbib19","doi-asserted-by":"publisher","first-page":"1348","DOI":"10.1364\/optica.3.001348","article-title":"Demonstration of a 4 \u00d7 4-port universal linear circuit","volume":"3","author":"Ribeiro","year":"2016","journal-title":"Optica"},{"key":"nceac724dbib20","doi-asserted-by":"publisher","first-page":"4501612","DOI":"10.1109\/jphot.2019.2952562","article-title":"A single layer neural network implemented by a 4 \u00d7 4 MZI-based optical processor","volume":"11","author":"Shokraneh","year":"2019","journal-title":"IEEE Photon. J."},{"key":"nceac724dbib21","doi-asserted-by":"publisher","DOI":"10.1103\/physrevapplied.11.064043","article-title":"Silicon photonic modulator neuron","volume":"11","author":"Tait","year":"2019","journal-title":"Phys. Rev. Appl."},{"key":"nceac724dbib22","doi-asserted-by":"publisher","first-page":"20456","DOI":"10.1364\/oe.27.020456","article-title":"Programmable matrix operation with reconfigurable time-wavelength plane manipulation and dispersed time delay","volume":"27","author":"Huang","year":"2019","journal-title":"Opt. Express"},{"key":"nceac724dbib23","doi-asserted-by":"publisher","first-page":"457","DOI":"10.1038\/s41467-020-20719-7","article-title":"An optical neural chip for implementing complex-valued neural network","volume":"12","author":"Zhang","year":"2021","journal-title":"Nat. Commun."},{"key":"nceac724dbib24","doi-asserted-by":"publisher","DOI":"10.1364\/ofc.2020.th4c.6","article-title":"Demonstration of photonic neural network for fiber nonlinearity compensation in long-haul transmission systems","author":"Huang","year":"2020"},{"key":"nceac724dbib25","doi-asserted-by":"publisher","DOI":"10.1364\/ofc.2020.m2k.4","article-title":"Real-time operation of silicon photonic neurons","author":"de Lima","year":"2020"},{"key":"nceac724dbib26","doi-asserted-by":"publisher","first-page":"7701111","DOI":"10.1109\/jstqe.2019.2945548","article-title":"Deep neural network through an InP SOA-based photonic integrated cross-connect","volume":"26","author":"Shi","year":"2020","journal-title":"IEEE J. Sel. Top. Quantum Electron."},{"key":"nceac724dbib27","article-title":"Microring weight banks control beyond 8.5-bits accuracy","author":"Zhang","year":"2021"},{"key":"nceac724dbib28","doi-asserted-by":"publisher","first-page":"52","DOI":"10.1038\/s41586-020-03070-1","article-title":"Parallel convolutional processing using an integrated photonic tensor core","volume":"589","author":"Feldmann","year":"2021","journal-title":"Nature"},{"key":"nceac724dbib29","doi-asserted-by":"publisher","first-page":"44","DOI":"10.1038\/s41586-020-03063-0","article-title":"11 TOPS photonic convolutional accelerator for optical neural networks","volume":"589","author":"Xu","year":"2021","journal-title":"Nature"},{"key":"nceac724dbib30","doi-asserted-by":"publisher","DOI":"10.1364\/ofc.2021.tu5h.1","article-title":"A silicon photonic coherent neuron with 10GMAC\/sec processing line-rate","author":"Mourgias-Alexandris","year":"2021"},{"key":"nceac724dbib31","first-page":"1","article-title":"Silicon-integrated coherent neurons with 32GMAC\/sec\/axon compute line-rates using EAM-based input and weighting cells","author":"Giamougiannis","year":"2021"},{"key":"nceac724dbib32","doi-asserted-by":"publisher","first-page":"8800115","DOI":"10.1109\/jstqe.2020.2975579","article-title":"Femtojoule per MAC neuromorphic photonics: an energy and technology roadmap","volume":"26","author":"Totovi\u0107","year":"2020","journal-title":"IEEE J. Sel. Top. Quantum Electron."},{"key":"nceac724dbib33","doi-asserted-by":"publisher","first-page":"7701518","DOI":"10.1109\/jstqe.2019.2941485","article-title":"Photonic multiply-accumulate operations for neural networks","volume":"26","author":"Nahmias","year":"2020","journal-title":"IEEE J. Sel. Top. Quantum Electron."},{"key":"nceac724dbib34","doi-asserted-by":"publisher","first-page":"58","DOI":"10.1103\/physrevlett.73.58","article-title":"Experimental realization of any discrete unitary operator","volume":"73","author":"Reck","year":"1994","journal-title":"Phys. Rev. Lett."},{"key":"nceac724dbib35","doi-asserted-by":"publisher","first-page":"1460","DOI":"10.1364\/optica.3.001460","article-title":"Optimal design for universal multiport interferometers","volume":"3","author":"Clements","year":"2016","journal-title":"Optica"},{"key":"nceac724dbib36","doi-asserted-by":"publisher","first-page":"811","DOI":"10.1109\/jlt.2019.2949133","article-title":"Neuromorphic photonics with coherent linear neurons using dual-IQ modulation cells","volume":"38","author":"Mourgias-Alexandris","year":"2019","journal-title":"J. Lightwave Technol."},{"key":"nceac724dbib37","article-title":"Coherent photonic crossbar as a universal linear operator","author":"Giamougiannis","year":"2021","journal-title":"Laser Photon. Rev."},{"key":"nceac724dbib38","doi-asserted-by":"publisher","first-page":"5605","DOI":"10.1038\/s41598-022-09370-y","article-title":"Programmable photonic neural networks combining WDM with coherent linear optics","volume":"12","author":"Totovic","year":"2022","journal-title":"Sci. Rep."},{"key":"nceac724dbib39","doi-asserted-by":"publisher","DOI":"10.1007\/s11432-020-3025-4","article-title":"Theoretical analysis of PAM-N and M-QAM BER computation with single-sideband signal","volume":"64","author":"Lu","year":"2021","journal-title":"Sci. China Inf. Sci."},{"key":"nceac724dbib40","doi-asserted-by":"publisher","first-page":"23495","DOI":"10.1364\/oe.395441","article-title":"The diamond mesh, a phase-error- and loss-tolerant field-programmable MZI-based optical processor for optical neural networks","volume":"28","author":"Shokraneh","year":"2020","journal-title":"Opt. Express"},{"key":"nceac724dbib41","doi-asserted-by":"publisher","first-page":"455","DOI":"10.1137\/07070111x","article-title":"Tensor decompositions and applications","volume":"51","author":"Kolda","year":"2009","journal-title":"SIAM Rev."},{"key":"nceac724dbib42","first-page":"1","article-title":"High performance tensor-vector multiplies on shared memory systems","author":"Pawlowski","year":"2019"},{"key":"nceac724dbib43","doi-asserted-by":"publisher","first-page":"7430","DOI":"10.1038\/s41598-017-07754-z","article-title":"Neuromorphic photonic networks using silicon photonic weight banks","volume":"7","author":"Tait","year":"2017","journal-title":"Sci. Rep."},{"key":"nceac724dbib44","doi-asserted-by":"publisher","first-page":"91","DOI":"10.1038\/s41377-020-0325-9","article-title":"Optical RAM and integrated optical memories: a survey","volume":"9","author":"Alexoudi","year":"2020","journal-title":"Light Sci. Appl."},{"key":"nceac724dbib45","doi-asserted-by":"publisher","DOI":"10.1063\/5.0001942","article-title":"Photonic tensor cores for machine learning","volume":"7","author":"Miscuglio","year":"2020","journal-title":"Appl. Phys. Rev."},{"key":"nceac724dbib46","doi-asserted-by":"publisher","first-page":"5140","DOI":"10.1109\/jlt.2019.2929593","article-title":"O-band silicon photonic transmitters for datacom and computercom interconnects","volume":"37","author":"Pitris","year":"2019","journal-title":"J. Lightwave Technol."},{"key":"nceac724dbib47","doi-asserted-by":"publisher","first-page":"3366","DOI":"10.1109\/jlt.2020.2977369","article-title":"400 Gb\/s silicon photonic transmitter and routing WDM technologies for glueless 8-socket chip-to-chip interconnects","volume":"38","author":"Pitris","year":"2020","journal-title":"J. Lightwave Technol."},{"key":"nceac724dbib48","doi-asserted-by":"publisher","first-page":"5706","DOI":"10.1364\/oe.373454","article-title":"4-channel 200 Gb\/s WDM O-band silicon photonic transceiver sub-assembly","volume":"28","author":"Moralis-Pegios","year":"2020","journal-title":"Opt. Express"},{"key":"nceac724dbib49","article-title":"The MNIST database of handwritten digits","author":"LeCun","year":"1998"},{"key":"nceac724dbib50","doi-asserted-by":"publisher","first-page":"2278","DOI":"10.1109\/5.726791","article-title":"Gradient-based learning applied to document recognition","volume":"86","author":"LeCun","year":"1998","journal-title":"Proc. IEEE"},{"key":"nceac724dbib51","author":"Goodfellow","year":"2016"},{"key":"nceac724dbib52","doi-asserted-by":"publisher","first-page":"9620","DOI":"10.1364\/oe.27.009620","article-title":"An all-optical neuron with sigmoid activation function","volume":"27","author":"Mourgias-Alexandris","year":"2019","journal-title":"Opt. Express"},{"key":"nceac724dbib53","doi-asserted-by":"publisher","first-page":"2003","DOI":"10.1364\/ol.422930","article-title":"Adaptive sigmoid-like and PReLU activation functions for all-optical perceptron","volume":"46","author":"Crnjanski","year":"2021","journal-title":"Opt. Lett."},{"key":"nceac724dbib54","author":"Hinton","year":"2012"},{"key":"nceac724dbib55","first-page":"8026","article-title":"PyTorch: an imperative style, high-performance deep learning library","volume":"vol 32","author":"Paszke","year":"2019"},{"key":"nceac724dbib56"}],"container-title":["Neuromorphic Computing and Engineering"],"original-title":[],"link":[{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2634-4386\/ac724d","content-type":"text\/html","content-version":"am","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2634-4386\/ac724d\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2634-4386\/ac724d","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2634-4386\/ac724d\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2634-4386\/ac724d\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2634-4386\/ac724d\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2634-4386\/ac724d\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"similarity-checking"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2634-4386\/ac724d\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2022,11,30]],"date-time":"2022-11-30T15:53:23Z","timestamp":1669823603000},"score":1,"resource":{"primary":{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2634-4386\/ac724d"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2022,6,1]]},"references-count":56,"journal-issue":{"issue":"2","published-online":{"date-parts":[[2022,6,20]]},"published-print":{"date-parts":[[2022,6,1]]}},"URL":"https:\/\/doi.org\/10.1088\/2634-4386\/ac724d","relation":{},"ISSN":["2634-4386"],"issn-type":[{"value":"2634-4386","type":"electronic"}],"subject":[],"published":{"date-parts":[[2022,6,1]]},"assertion":[{"value":"WDM equipped universal linear optics for programmable neuromorphic photonic processors","name":"article_title","label":"Article Title"},{"value":"Neuromorphic Computing and Engineering","name":"journal_title","label":"Journal Title"},{"value":"paper","name":"article_type","label":"Article Type"},{"value":"\u00a9 2022 The Author(s). Published by IOP Publishing Ltd","name":"copyright_information","label":"Copyright Information"},{"value":"2021-12-31","name":"date_received","label":"Date Received","group":{"name":"publication_dates","label":"Publication dates"}},{"value":"2022-05-23","name":"date_accepted","label":"Date Accepted","group":{"name":"publication_dates","label":"Publication dates"}},{"value":"2022-06-20","name":"date_epub","label":"Online publication date","group":{"name":"publication_dates","label":"Publication dates"}}]}}