Matches in SemOpenAlex for { <https://semopenalex.org/work/W3038313908> ?p ?o ?g. }
- W3038313908 abstract "Neuromorphic computing with analog memory can accelerate deep neural networks (DNNs) by enabling multiply-accumulate (MAC) operations to occur within memory. Analog memory, however, presents a number of device-level challenges having macro-implications on the achievable accuracy and reliability of these artificial neural networks. This paper focuses on the adverse effects of conductance drift in phase-change memory (PCM) on network reliability. It is shown that conductance drift can be effectively compensated in a variety of networks by applying a ‘slope correction’ technique to the squashing functions to maintain accuracy/reliability for a period of ~1 year. In addition to conductance drift, PCM poses considerable variability challenges, which impact the accuracy of the initial t 0 weights. This paper summarizes recent advances in optimizing t 0 weight programming, and provides evidence suggesting that the combination of ‘slope correction’ and programming optimization techniques may allow DNN acceleration using analog memory while maintaining software-equivalent accuracy with reasonable reliability." @default.
- W3038313908 created "2020-07-10" @default.
- W3038313908 creator A5006344168 @default.
- W3038313908 creator A5014981734 @default.
- W3038313908 creator A5015811754 @default.
- W3038313908 creator A5026800272 @default.
- W3038313908 creator A5028455655 @default.
- W3038313908 creator A5035067308 @default.
- W3038313908 creator A5042571447 @default.
- W3038313908 creator A5058290381 @default.
- W3038313908 creator A5079173979 @default.
- W3038313908 creator A5086723230 @default.
- W3038313908 date "2020-04-01" @default.
- W3038313908 modified "2023-10-14" @default.
- W3038313908 title "Neuromorphic Computing with Phase Change, Device Reliability, and Variability Challenges" @default.
- W3038313908 cites W1500058340 @default.
- W3038313908 cites W1655377202 @default.
- W3038313908 cites W1677182931 @default.
- W3038313908 cites W1937359183 @default.
- W3038313908 cites W1989174502 @default.
- W3038313908 cites W1997785184 @default.
- W3038313908 cites W2032241398 @default.
- W3038313908 cites W2036481447 @default.
- W3038313908 cites W2060501200 @default.
- W3038313908 cites W2070090854 @default.
- W3038313908 cites W2074452321 @default.
- W3038313908 cites W2081381639 @default.
- W3038313908 cites W2120286148 @default.
- W3038313908 cites W2120432001 @default.
- W3038313908 cites W2122352981 @default.
- W3038313908 cites W2126010728 @default.
- W3038313908 cites W2158464944 @default.
- W3038313908 cites W2162811812 @default.
- W3038313908 cites W2257979135 @default.
- W3038313908 cites W2433234826 @default.
- W3038313908 cites W2525141714 @default.
- W3038313908 cites W2531533975 @default.
- W3038313908 cites W2539625543 @default.
- W3038313908 cites W2560615381 @default.
- W3038313908 cites W2621123884 @default.
- W3038313908 cites W2740220207 @default.
- W3038313908 cites W2741948058 @default.
- W3038313908 cites W2743575928 @default.
- W3038313908 cites W2743999796 @default.
- W3038313908 cites W2769049661 @default.
- W3038313908 cites W2778345336 @default.
- W3038313908 cites W2785346409 @default.
- W3038313908 cites W2785784536 @default.
- W3038313908 cites W2807813675 @default.
- W3038313908 cites W2886186645 @default.
- W3038313908 cites W2893064244 @default.
- W3038313908 cites W2910983724 @default.
- W3038313908 cites W2919115771 @default.
- W3038313908 cites W2942077180 @default.
- W3038313908 cites W2964447465 @default.
- W3038313908 cites W3005874416 @default.
- W3038313908 doi "https://doi.org/10.1109/irps45951.2020.9128315" @default.
- W3038313908 hasPublicationYear "2020" @default.
- W3038313908 type Work @default.
- W3038313908 sameAs 3038313908 @default.
- W3038313908 citedByCount "3" @default.
- W3038313908 countsByYear W30383139082021 @default.
- W3038313908 crossrefType "proceedings-article" @default.
- W3038313908 hasAuthorship W3038313908A5006344168 @default.
- W3038313908 hasAuthorship W3038313908A5014981734 @default.
- W3038313908 hasAuthorship W3038313908A5015811754 @default.
- W3038313908 hasAuthorship W3038313908A5026800272 @default.
- W3038313908 hasAuthorship W3038313908A5028455655 @default.
- W3038313908 hasAuthorship W3038313908A5035067308 @default.
- W3038313908 hasAuthorship W3038313908A5042571447 @default.
- W3038313908 hasAuthorship W3038313908A5058290381 @default.
- W3038313908 hasAuthorship W3038313908A5079173979 @default.
- W3038313908 hasAuthorship W3038313908A5086723230 @default.
- W3038313908 hasConcept C113775141 @default.
- W3038313908 hasConcept C117896860 @default.
- W3038313908 hasConcept C121332964 @default.
- W3038313908 hasConcept C127413603 @default.
- W3038313908 hasConcept C133256868 @default.
- W3038313908 hasConcept C151927369 @default.
- W3038313908 hasConcept C154945302 @default.
- W3038313908 hasConcept C163258240 @default.
- W3038313908 hasConcept C41008148 @default.
- W3038313908 hasConcept C43214815 @default.
- W3038313908 hasConcept C50644808 @default.
- W3038313908 hasConcept C61696701 @default.
- W3038313908 hasConcept C62520636 @default.
- W3038313908 hasConcept C64142963 @default.
- W3038313908 hasConcept C74650414 @default.
- W3038313908 hasConceptScore W3038313908C113775141 @default.
- W3038313908 hasConceptScore W3038313908C117896860 @default.
- W3038313908 hasConceptScore W3038313908C121332964 @default.
- W3038313908 hasConceptScore W3038313908C127413603 @default.
- W3038313908 hasConceptScore W3038313908C133256868 @default.
- W3038313908 hasConceptScore W3038313908C151927369 @default.
- W3038313908 hasConceptScore W3038313908C154945302 @default.
- W3038313908 hasConceptScore W3038313908C163258240 @default.
- W3038313908 hasConceptScore W3038313908C41008148 @default.
- W3038313908 hasConceptScore W3038313908C43214815 @default.
- W3038313908 hasConceptScore W3038313908C50644808 @default.
- W3038313908 hasConceptScore W3038313908C61696701 @default.