Artykuł w czasopiśmie
Brak miniatury
Licencja
ClosedAccessDostęp zamknięty
 

Energy-Efficient Neural Network Inference with Microcavity Exciton Polaritons

Uproszczony widok
dc.abstract.enWe propose all-optical neural networks characterized by very high energy efficiency and performance density of inference. We argue that the use of microcavity exciton polaritons allows one to take advantage of the properties of both photons and electrons in a seamless manner. This results in strong optical nonlinearity without the use of optoelectronic conversion. We propose a design of a realistic neural network and estimate energy cost to be at the level of attojoules per bit, also when including the optoelectronic conversion at the input and output of the network, several orders of magnitude below state-of-the-art hardware implementations. We propose two kinds of nonlinear
dc.affiliationUniwersytet Warszawski
dc.contributor.authorFurman, Magdalena
dc.contributor.authorKról, Marcin
dc.contributor.authorTyszka, Krzysztof
dc.contributor.authorMirek, Rafał
dc.contributor.authorOpala, Andrzej
dc.contributor.authorMatuszewski, Michał
dc.contributor.authorSanvitto, D.
dc.contributor.authorBallarini, D.
dc.contributor.authorLiew, T.C.H.
dc.contributor.authorSzczytko, Jacek
dc.contributor.authorPiętka, Barbara
dc.date.accessioned2024-01-24T22:47:44Z
dc.date.available2024-01-24T22:47:44Z
dc.date.issued2021
dc.description.financePublikacja bezkosztowa
dc.description.number2
dc.description.volume16
dc.identifier.doi10.1103/PHYSREVAPPLIED.16.024045
dc.identifier.urihttps://repozytorium.uw.edu.pl//handle/item/106176
dc.identifier.weblinkhttps://link.aps.org/article/10.1103/PhysRevApplied.16.024045
dc.languageeng
dc.pbn.affiliationphysical sciences
dc.relation.ispartofPhysical Review Applied
dc.relation.pages024045-1-16
dc.rightsClosedAccess
dc.sciencecloudnosend
dc.titleEnergy-Efficient Neural Network Inference with Microcavity Exciton Polaritons
dc.typeJournalArticle
dspace.entity.typePublication