Matches in SemOpenAlex for { <https://semopenalex.org/work/W2904694042> ?p ?o ?g. }
- W2904694042 abstract "Recent success in deep neural networks has generated strong interest in hardware accelerators to improve speed and energy consumption. This paper presents a new type of photonic accelerator based on coherent detection that is scalable to large ($N gtrsim 10^6$) networks and can be operated at high (GHz) speeds and very low (sub-aJ) energies per multiply-and-accumulate (MAC), using the massive spatial multiplexing enabled by standard free-space optical components. In contrast to previous approaches, both weights and inputs are optically encoded so that the network can be reprogrammed and trained on the fly. Simulations of the network using models for digit- and image-classification reveal a standard quantum limit for optical neural networks, set by photodetector shot noise. This bound, which can be as low as 50 zJ/MAC, suggests performance below the thermodynamic (Landauer) limit for digital irreversible computation is theoretically possible in this device. The proposed accelerator can implement both fully-connected and convolutional networks. We also present a scheme for back-propagation and training that can be performed in the same hardware. This architecture will enable a new class of ultra-low-energy processors for deep learning." @default.
- W2904694042 created "2018-12-22" @default.
- W2904694042 creator A5008843520 @default.
- W2904694042 creator A5027186207 @default.
- W2904694042 creator A5049120533 @default.
- W2904694042 creator A5060426875 @default.
- W2904694042 creator A5089821617 @default.
- W2904694042 date "2019-05-16" @default.
- W2904694042 modified "2023-10-16" @default.
- W2904694042 title "Large-Scale Optical Neural Networks Based on Photoelectric Multiplication" @default.
- W2904694042 cites W1971094900 @default.
- W2904694042 cites W1978718464 @default.
- W2904694042 cites W1988425770 @default.
- W2904694042 cites W1989478505 @default.
- W2904694042 cites W1996408406 @default.
- W2904694042 cites W2005886497 @default.
- W2904694042 cites W2010273331 @default.
- W2904694042 cites W2012289238 @default.
- W2904694042 cites W2012683031 @default.
- W2904694042 cites W2021379807 @default.
- W2904694042 cites W2022248835 @default.
- W2904694042 cites W2024122052 @default.
- W2904694042 cites W2029287805 @default.
- W2904694042 cites W2031788394 @default.
- W2904694042 cites W2040870580 @default.
- W2904694042 cites W2042545232 @default.
- W2904694042 cites W2043163981 @default.
- W2904694042 cites W2051270432 @default.
- W2904694042 cites W2054106534 @default.
- W2904694042 cites W2062294208 @default.
- W2904694042 cites W2070225719 @default.
- W2904694042 cites W2076476236 @default.
- W2904694042 cites W2098874108 @default.
- W2904694042 cites W2099108362 @default.
- W2904694042 cites W2108491246 @default.
- W2904694042 cites W2110427327 @default.
- W2904694042 cites W2111721632 @default.
- W2904694042 cites W2112796928 @default.
- W2904694042 cites W2117539524 @default.
- W2904694042 cites W2124072147 @default.
- W2904694042 cites W2128960304 @default.
- W2904694042 cites W2152839228 @default.
- W2904694042 cites W2161376274 @default.
- W2904694042 cites W2220870060 @default.
- W2904694042 cites W2289252105 @default.
- W2904694042 cites W2314444502 @default.
- W2904694042 cites W2323986115 @default.
- W2904694042 cites W2324271557 @default.
- W2904694042 cites W2338018404 @default.
- W2904694042 cites W2343850787 @default.
- W2904694042 cites W2345430977 @default.
- W2904694042 cites W2428087307 @default.
- W2904694042 cites W2538025098 @default.
- W2904694042 cites W2565275310 @default.
- W2904694042 cites W2604319603 @default.
- W2904694042 cites W2623643194 @default.
- W2904694042 cites W2736319828 @default.
- W2904694042 cites W2739588406 @default.
- W2904694042 cites W2752849906 @default.
- W2904694042 cites W2766733243 @default.
- W2904694042 cites W2797501353 @default.
- W2904694042 cites W2798125601 @default.
- W2904694042 cites W2798701005 @default.
- W2904694042 cites W2807750997 @default.
- W2904694042 cites W2884001105 @default.
- W2904694042 cites W2887566712 @default.
- W2904694042 cites W2899558822 @default.
- W2904694042 cites W2902907165 @default.
- W2904694042 cites W2962730419 @default.
- W2904694042 cites W2963912275 @default.
- W2904694042 cites W3101465594 @default.
- W2904694042 cites W3103448394 @default.
- W2904694042 cites W3194987791 @default.
- W2904694042 cites W4233798822 @default.
- W2904694042 doi "https://doi.org/10.1103/physrevx.9.021032" @default.
- W2904694042 hasPublicationYear "2019" @default.
- W2904694042 type Work @default.
- W2904694042 sameAs 2904694042 @default.
- W2904694042 citedByCount "154" @default.
- W2904694042 countsByYear W29046940422019 @default.
- W2904694042 countsByYear W29046940422020 @default.
- W2904694042 countsByYear W29046940422021 @default.
- W2904694042 countsByYear W29046940422022 @default.
- W2904694042 countsByYear W29046940422023 @default.
- W2904694042 crossrefType "journal-article" @default.
- W2904694042 hasAuthorship W2904694042A5008843520 @default.
- W2904694042 hasAuthorship W2904694042A5027186207 @default.
- W2904694042 hasAuthorship W2904694042A5049120533 @default.
- W2904694042 hasAuthorship W2904694042A5060426875 @default.
- W2904694042 hasAuthorship W2904694042A5089821617 @default.
- W2904694042 hasBestOaLocation W29046940421 @default.
- W2904694042 hasConcept C108583219 @default.
- W2904694042 hasConcept C113775141 @default.
- W2904694042 hasConcept C115961682 @default.
- W2904694042 hasConcept C120665830 @default.
- W2904694042 hasConcept C121332964 @default.
- W2904694042 hasConcept C13164978 @default.
- W2904694042 hasConcept C134306372 @default.
- W2904694042 hasConcept C151201525 @default.
- W2904694042 hasConcept C154945302 @default.