Consiglio Nazionale delle Ricerche

Tipo di prodottoArticolo in rivista
TitoloTheory of Neuromorphic Computing by Waves: Machine Learning by Rogue Waves, Dispersive Shocks, and Solitons
Anno di pubblicazione2020
FormatoElettronico
Autore/iMarcucci G.; Pierangeli D.; Conti C.
Affiliazioni autoriInstitute for Complex Systems, National Research Council (ISC-CNR), Italy and Department of Physics, Sapienza University, Via dei Taurini 19Piazzale Aldo Moro 2, Rome, 00185, Italy; Institute for Complex Systems, National Research Council (ISC-CNR), Italy and Department of Physics, Sapienza University, Via dei Taurini 19Piazzale Aldo Moro 2, Rome, 00185, Italy; Institute for Complex Systems, National Research Council (ISC-CNR), Italy and Department of Physics, Sapienza University, Via dei Taurini 19Piazzale Aldo Moro 2, Rome, 00185, Italy
Autori CNR e affiliazioni
  • GIULIA MARCUCCI
  • DAVIDE PIERANGELI
  • CLAUDIO CONTI
Lingua/e
  • inglese
AbstractWe study artificial neural networks with nonlinear waves as a computing reservoir. We discuss universality and the conditions to learn a dataset in terms of output channels and nonlinearity. A feed-forward three-layered model, with an encoding input layer, a wave layer, and a decoding readout, behaves as a conventional neural network in approximating mathematical functions, real-world datasets, and universal Boolean gates. The rank of the transmission matrix has a fundamental role in assessing the learning abilities of the wave. For a given set of training points, a threshold nonlinearity for universal interpolation exists. When considering the nonlinear Schrödinger equation, the use of highly nonlinear regimes implies that solitons, rogue, and shock waves do have a leading role in training and computing. Our results may enable the realization of novel machine learning devices by using diverse physical systems, as nonlinear optics, hydrodynamics, polaritonics, and Bose-Einstein condensates. The application of these concepts to photonics opens the way to a large class of accelerators and new computational paradigms. In complex wave systems, as multimodal fibers, integrated optical circuits, random, topological devices, and metasurfaces, nonlinear waves can be employed to perform computation and solve complex combinatorial optimization.
Lingua abstractinglese
Altro abstract-
Lingua altro abstract-
Pagine da093901
Pagine a-
Pagine totali6
RivistaPhysical review letters
Attiva dal 1958
Editore: American Physical Society - College Park, MD
Paese di pubblicazione: Stati Uniti d'America
Lingua: inglese
ISSN: 1079-7114
Titolo chiave: Physical review letters
Titolo proprio: Physical review letters
Titolo abbreviato: Phys. rev. letters
Titoli alternativi:
  • Physical review letters online
  • PRL
Numero volume della rivista125
Fascicolo della rivista9
DOI10.1103/PhysRevLett.125.093901
Verificato da referee-
Stato della pubblicazionePublished version
Indicizzazione (in banche dati controllate)
  • Scopus (Codice:2-s2.0-85090914417)
  • ISI Web of Science (WOS) (Codice:000562633900005)
Parole chiaveNETWORKS
Link (URL, URI)https://journals.aps.org/prl/abstract/10.1103/PhysRevLett.125.093901
Titolo parallelo-
Licenza-
Scadenza embargo-
Data di accettazione09/07/2020
Note/Altre informazioniFunding: QuantERA ERA-NET Co-fund (Grant No. 731473,QUOMPLEX) PRIN PELM (20177PSCKT) H2020 PhoQus (Grant No. 820392)
Strutture CNR
  • ISC — Istituto dei sistemi complessi
Moduli/Attività/Sottoprogetti CNR-
Progetti Europei-
Allegati
Theory of Neuromorphic Computing by Waves: Machine Learning by Rogue Waves, Dispersive Shocks, and Solitons (documento privato )
Descrizione: published version
Tipo documento: application/pdf