Matches in SemOpenAlex for { <https://semopenalex.org/work/W2968714528> ?p ?o ?g. }
- W2968714528 endingPage "543" @default.
- W2968714528 startingPage "532" @default.
- W2968714528 abstract "The second wave of AI is about statistical learning of low dimensional structures from high dimensional data. Inference is done using multilayer, data transforming networks using fixed point arithmetic with parameters that have limited precision (4-16 bits). In this paper we give a historical perspective on hardware AI inference deep Artificial Neural Networks (ANNs) or in short Deep Neural Networks (DNN) and deep learning. We review custom chip implementations of ANNs from thirty years ago and from the more recent publications in the last five years. With only few exceptions, hardware AI architectures are digital but we argue that if done right, i.e. in the charge domain, analog computation will have a role in future hardware AI systems. We make our discussion concrete by presenting the architecture, implementation and measurements from a mixedsignal, charge-based 8-bit analog multiplier for limited precision linear algebra in AI systems. Using a capacitor array, and charge redistribution, the architecture performs the multiplication operation in the charge domain at the thermal noise limit with near minimum energy dissipation. The charge redistribution multiplier core was fabricated in a 16 nm FinFET CMOS process, with measured energy 1.4 fJ for the analog multiplication operation. Compared to a conventional digital implementation synthesized and simulated in the same technology, the proposed design achieves the same performance at 37% less energy." @default.
- W2968714528 created "2019-08-22" @default.
- W2968714528 creator A5057789499 @default.
- W2968714528 creator A5082255625 @default.
- W2968714528 date "2019-09-01" @default.
- W2968714528 modified "2023-10-01" @default.
- W2968714528 title "A Historical Perspective on Hardware AI Inference, Charge-Based Computational Circuits and an 8 bit Charge-Based Multiply-Add Core in 16 nm FinFET CMOS" @default.
- W2968714528 cites W1637630365 @default.
- W2968714528 cites W1851583682 @default.
- W2968714528 cites W1964559842 @default.
- W2968714528 cites W1968422655 @default.
- W2968714528 cites W1980661496 @default.
- W2968714528 cites W1981071268 @default.
- W2968714528 cites W1986558808 @default.
- W2968714528 cites W1987327272 @default.
- W2968714528 cites W1989475960 @default.
- W2968714528 cites W1996821523 @default.
- W2968714528 cites W1997960524 @default.
- W2968714528 cites W1999497791 @default.
- W2968714528 cites W2017540033 @default.
- W2968714528 cites W2019138730 @default.
- W2968714528 cites W2020376324 @default.
- W2968714528 cites W2022983552 @default.
- W2968714528 cites W2024122052 @default.
- W2968714528 cites W2025201350 @default.
- W2968714528 cites W2040870580 @default.
- W2968714528 cites W2041141748 @default.
- W2968714528 cites W2044437287 @default.
- W2968714528 cites W2057227450 @default.
- W2968714528 cites W2066493139 @default.
- W2968714528 cites W2072231431 @default.
- W2968714528 cites W2080504155 @default.
- W2968714528 cites W2091006540 @default.
- W2968714528 cites W2101586548 @default.
- W2968714528 cites W2112796928 @default.
- W2968714528 cites W2118814875 @default.
- W2968714528 cites W2119397956 @default.
- W2968714528 cites W2119732736 @default.
- W2968714528 cites W2123418793 @default.
- W2968714528 cites W2125010820 @default.
- W2968714528 cites W2132131403 @default.
- W2968714528 cites W2132406028 @default.
- W2968714528 cites W2132975313 @default.
- W2968714528 cites W2137893918 @default.
- W2968714528 cites W2142037882 @default.
- W2968714528 cites W2142753319 @default.
- W2968714528 cites W2142850400 @default.
- W2968714528 cites W2144798795 @default.
- W2968714528 cites W2147220807 @default.
- W2968714528 cites W2161028202 @default.
- W2968714528 cites W2163630896 @default.
- W2968714528 cites W2289252105 @default.
- W2968714528 cites W2334805829 @default.
- W2968714528 cites W2340076492 @default.
- W2968714528 cites W2509538613 @default.
- W2968714528 cites W2550848904 @default.
- W2968714528 cites W2587415333 @default.
- W2968714528 cites W2606722458 @default.
- W2968714528 cites W2613142224 @default.
- W2968714528 cites W2613967272 @default.
- W2968714528 cites W2622148586 @default.
- W2968714528 cites W2625200202 @default.
- W2968714528 cites W2787779756 @default.
- W2968714528 cites W2792607576 @default.
- W2968714528 cites W2801611652 @default.
- W2968714528 cites W2898985762 @default.
- W2968714528 cites W2898991608 @default.
- W2968714528 cites W2899481200 @default.
- W2968714528 cites W2907232526 @default.
- W2968714528 cites W2919115771 @default.
- W2968714528 cites W2920326572 @default.
- W2968714528 cites W4247114301 @default.
- W2968714528 cites W4253020087 @default.
- W2968714528 doi "https://doi.org/10.1109/jetcas.2019.2933795" @default.
- W2968714528 hasPublicationYear "2019" @default.
- W2968714528 type Work @default.
- W2968714528 sameAs 2968714528 @default.
- W2968714528 citedByCount "8" @default.
- W2968714528 countsByYear W29687145282020 @default.
- W2968714528 countsByYear W29687145282021 @default.
- W2968714528 countsByYear W29687145282022 @default.
- W2968714528 countsByYear W29687145282023 @default.
- W2968714528 crossrefType "journal-article" @default.
- W2968714528 hasAuthorship W2968714528A5057789499 @default.
- W2968714528 hasAuthorship W2968714528A5082255625 @default.
- W2968714528 hasConcept C113775141 @default.
- W2968714528 hasConcept C11413529 @default.
- W2968714528 hasConcept C121332964 @default.
- W2968714528 hasConcept C124584101 @default.
- W2968714528 hasConcept C127413603 @default.
- W2968714528 hasConcept C135402231 @default.
- W2968714528 hasConcept C139719470 @default.
- W2968714528 hasConcept C154945302 @default.
- W2968714528 hasConcept C162324750 @default.
- W2968714528 hasConcept C24326235 @default.
- W2968714528 hasConcept C41008148 @default.
- W2968714528 hasConcept C459310 @default.
- W2968714528 hasConcept C46362747 @default.