Matches in SemOpenAlex for { <https://semopenalex.org/work/W4285178986> ?p ?o ?g. }
- W4285178986 endingPage "643" @default.
- W4285178986 startingPage "630" @default.
- W4285178986 abstract "Stochastic computing (SC) has become a promising approximate computing solution by its negligible resource occupancy and ultralow energy consumption. As a potential replacement of accurate multiplication, SC can dramatically mitigate the problematic power consumption by DNNs. However, current SC-multipliers illustrate an extremely imbalanced accuracy across product space, i.e., neglectable noise with large products but significant noise for small ones, which is discordant to the distribution of products by the sparse matrix in neural computing. In this article, we present a heterogeneous SC-multiplier that heuristically performs three divergent approximating multiplication, including “set-to-0,” “look-up-table,” and “low-discrepancy-SC,” for appropriate precision-provision in the whole space of products. Due to those popular DNN models cannot achieve consensus on the boundaries of above operations, a training-involved method is proposed to determine the settings with limited overhead. In this way, those models successively learn the SC-operation characters and exhibit a definitely improvement on network precision. The experiment shows that, for single multiplication, the product noise can be restrained by 36.86% on average, and for multiplication in multiple network models, the accuracy improvement reaches to 5.5% on average. Furthermore, a group of proposed logic-reduction techniques can improve the energy efficiency by 65% in the system-level evaluation." @default.
- W4285178986 created "2022-07-14" @default.
- W4285178986 creator A5003409994 @default.
- W4285178986 creator A5017058932 @default.
- W4285178986 creator A5022499603 @default.
- W4285178986 creator A5026165169 @default.
- W4285178986 creator A5034840700 @default.
- W4285178986 creator A5060964806 @default.
- W4285178986 date "2023-02-01" @default.
- W4285178986 modified "2023-10-09" @default.
- W4285178986 title "A Noise-Driven Heterogeneous Stochastic Computing Multiplier for Heuristic Precision Improvement in Energy-Efficient DNNs" @default.
- W4285178986 cites W2024203951 @default.
- W4285178986 cites W2067121838 @default.
- W4285178986 cites W2077344647 @default.
- W4285178986 cites W2098470668 @default.
- W4285178986 cites W2533474659 @default.
- W4285178986 cites W2545625532 @default.
- W4285178986 cites W2551814622 @default.
- W4285178986 cites W2579041139 @default.
- W4285178986 cites W2589986623 @default.
- W4285178986 cites W2614053718 @default.
- W4285178986 cites W2625226188 @default.
- W4285178986 cites W2802559652 @default.
- W4285178986 cites W2908853553 @default.
- W4285178986 cites W2963196212 @default.
- W4285178986 cites W2982479999 @default.
- W4285178986 cites W2982668311 @default.
- W4285178986 cites W3001688831 @default.
- W4285178986 cites W3003610858 @default.
- W4285178986 cites W3083027615 @default.
- W4285178986 cites W3089531544 @default.
- W4285178986 cites W3092331354 @default.
- W4285178986 cites W3126267573 @default.
- W4285178986 cites W3127007112 @default.
- W4285178986 cites W3164551710 @default.
- W4285178986 cites W3176350758 @default.
- W4285178986 cites W3177113534 @default.
- W4285178986 cites W4200387195 @default.
- W4285178986 cites W4200437691 @default.
- W4285178986 doi "https://doi.org/10.1109/tcad.2022.3178053" @default.
- W4285178986 hasPublicationYear "2023" @default.
- W4285178986 type Work @default.
- W4285178986 citedByCount "1" @default.
- W4285178986 countsByYear W42851789862023 @default.
- W4285178986 crossrefType "journal-article" @default.
- W4285178986 hasAuthorship W4285178986A5003409994 @default.
- W4285178986 hasAuthorship W4285178986A5017058932 @default.
- W4285178986 hasAuthorship W4285178986A5022499603 @default.
- W4285178986 hasAuthorship W4285178986A5026165169 @default.
- W4285178986 hasAuthorship W4285178986A5034840700 @default.
- W4285178986 hasAuthorship W4285178986A5060964806 @default.
- W4285178986 hasConcept C111919701 @default.
- W4285178986 hasConcept C113775141 @default.
- W4285178986 hasConcept C11413529 @default.
- W4285178986 hasConcept C114614502 @default.
- W4285178986 hasConcept C115961682 @default.
- W4285178986 hasConcept C119599485 @default.
- W4285178986 hasConcept C121332964 @default.
- W4285178986 hasConcept C124584101 @default.
- W4285178986 hasConcept C127413603 @default.
- W4285178986 hasConcept C134835016 @default.
- W4285178986 hasConcept C139719470 @default.
- W4285178986 hasConcept C154945302 @default.
- W4285178986 hasConcept C162324750 @default.
- W4285178986 hasConcept C163294075 @default.
- W4285178986 hasConcept C17349429 @default.
- W4285178986 hasConcept C173801870 @default.
- W4285178986 hasConcept C18903297 @default.
- W4285178986 hasConcept C199360897 @default.
- W4285178986 hasConcept C2742236 @default.
- W4285178986 hasConcept C2779960059 @default.
- W4285178986 hasConcept C2780165032 @default.
- W4285178986 hasConcept C2780595030 @default.
- W4285178986 hasConcept C2780971903 @default.
- W4285178986 hasConcept C33923547 @default.
- W4285178986 hasConcept C41008148 @default.
- W4285178986 hasConcept C50644808 @default.
- W4285178986 hasConcept C62520636 @default.
- W4285178986 hasConcept C84114770 @default.
- W4285178986 hasConcept C86803240 @default.
- W4285178986 hasConcept C99498987 @default.
- W4285178986 hasConceptScore W4285178986C111919701 @default.
- W4285178986 hasConceptScore W4285178986C113775141 @default.
- W4285178986 hasConceptScore W4285178986C11413529 @default.
- W4285178986 hasConceptScore W4285178986C114614502 @default.
- W4285178986 hasConceptScore W4285178986C115961682 @default.
- W4285178986 hasConceptScore W4285178986C119599485 @default.
- W4285178986 hasConceptScore W4285178986C121332964 @default.
- W4285178986 hasConceptScore W4285178986C124584101 @default.
- W4285178986 hasConceptScore W4285178986C127413603 @default.
- W4285178986 hasConceptScore W4285178986C134835016 @default.
- W4285178986 hasConceptScore W4285178986C139719470 @default.
- W4285178986 hasConceptScore W4285178986C154945302 @default.
- W4285178986 hasConceptScore W4285178986C162324750 @default.
- W4285178986 hasConceptScore W4285178986C163294075 @default.
- W4285178986 hasConceptScore W4285178986C17349429 @default.
- W4285178986 hasConceptScore W4285178986C173801870 @default.
- W4285178986 hasConceptScore W4285178986C18903297 @default.