Matches in SemOpenAlex for { <https://semopenalex.org/work/W2005536355> ?p ?o ?g. }
Showing items 1 to 97 of
97
with 100 items per page.
- W2005536355 endingPage "21" @default.
- W2005536355 startingPage "15" @default.
- W2005536355 abstract "Low overhead error-resiliency techniques such as RAZOR [1] and algorithmic noise-tolerance (ANT) [2] have proven effective in reducing energy consumption. ANT has been shown to be particularly effective for signal processing and machine learning kernels. In ANT, an explicit estimator block compensates for large magnitude errors in a main block. The estimator represents the overhead in ANT and can be as large as 30%. This paper presents a low overhead ANT technique referred to as ALG-ANT. In ALG-ANT, the estimator is embedded inside the main block via algorithmic reformulation and thus completely eliminates the overhead associated with ANT. However, ALG-ANT is algorithm-specific. This paper demonstrates the ALG-ANT concept in the context of a finite impulse response (FIR) filter kernel and a dot product kernel, both of which are commonly employed in signal processing and machine learning applications. The proposed ALG-ANT FIR filter and dot product kernels are applied to the feature extractor (FE) and SVM classification engine (CE) of an EEG seizure classification system. Simulation results in a commercial 45nm CMOS process show that ALG-ANT can compensate for error rates of up to 0.41 (errors in FE only), and up to 0.19 (errors in FE and CE) and maintain the true positive rate ptp > 0.9 and false positive rate pfp ≤ 0.01. This represents a greater than 3-orders-of-magnitude improvement in error tolerance over the conventional architecture. This error tolerance is employed to reduce energy via the use of voltage overscaling (VOS). ALG-ANT is able to achieve 44.3% energy savings when errors are in FE only, and up to 37.1% savings when errors are in both FE and CE." @default.
- W2005536355 created "2016-06-24" @default.
- W2005536355 creator A5032264417 @default.
- W2005536355 creator A5057014407 @default.
- W2005536355 date "2015-11-02" @default.
- W2005536355 modified "2023-09-26" @default.
- W2005536355 title "Reduced Overhead Error Compensation for Energy Efficient Machine Learning Kernels" @default.
- W2005536355 cites W123426918 @default.
- W2005536355 cites W1766888123 @default.
- W2005536355 cites W1973423687 @default.
- W2005536355 cites W1980246793 @default.
- W2005536355 cites W2010468809 @default.
- W2005536355 cites W2018493976 @default.
- W2005536355 cites W2023453288 @default.
- W2005536355 cites W2031547998 @default.
- W2005536355 cites W2042682017 @default.
- W2005536355 cites W2057558141 @default.
- W2005536355 cites W2102545063 @default.
- W2005536355 cites W2104677471 @default.
- W2005536355 cites W2113145298 @default.
- W2005536355 cites W2115722171 @default.
- W2005536355 cites W2117648153 @default.
- W2005536355 cites W2118331018 @default.
- W2005536355 cites W2149298154 @default.
- W2005536355 cites W2151802820 @default.
- W2005536355 cites W2160929292 @default.
- W2005536355 cites W2161969046 @default.
- W2005536355 cites W2296253867 @default.
- W2005536355 doi "https://doi.org/10.5555/2840819.2840822" @default.
- W2005536355 hasPublicationYear "2015" @default.
- W2005536355 type Work @default.
- W2005536355 sameAs 2005536355 @default.
- W2005536355 citedByCount "3" @default.
- W2005536355 countsByYear W20055363552016 @default.
- W2005536355 countsByYear W20055363552020 @default.
- W2005536355 crossrefType "proceedings-article" @default.
- W2005536355 hasAuthorship W2005536355A5032264417 @default.
- W2005536355 hasAuthorship W2005536355A5057014407 @default.
- W2005536355 hasConcept C105795698 @default.
- W2005536355 hasConcept C111919701 @default.
- W2005536355 hasConcept C11413529 @default.
- W2005536355 hasConcept C114614502 @default.
- W2005536355 hasConcept C154945302 @default.
- W2005536355 hasConcept C185429906 @default.
- W2005536355 hasConcept C198386975 @default.
- W2005536355 hasConcept C2524010 @default.
- W2005536355 hasConcept C2777210771 @default.
- W2005536355 hasConcept C2779960059 @default.
- W2005536355 hasConcept C33923547 @default.
- W2005536355 hasConcept C40128228 @default.
- W2005536355 hasConcept C40969351 @default.
- W2005536355 hasConcept C41008148 @default.
- W2005536355 hasConcept C74193536 @default.
- W2005536355 hasConceptScore W2005536355C105795698 @default.
- W2005536355 hasConceptScore W2005536355C111919701 @default.
- W2005536355 hasConceptScore W2005536355C11413529 @default.
- W2005536355 hasConceptScore W2005536355C114614502 @default.
- W2005536355 hasConceptScore W2005536355C154945302 @default.
- W2005536355 hasConceptScore W2005536355C185429906 @default.
- W2005536355 hasConceptScore W2005536355C198386975 @default.
- W2005536355 hasConceptScore W2005536355C2524010 @default.
- W2005536355 hasConceptScore W2005536355C2777210771 @default.
- W2005536355 hasConceptScore W2005536355C2779960059 @default.
- W2005536355 hasConceptScore W2005536355C33923547 @default.
- W2005536355 hasConceptScore W2005536355C40128228 @default.
- W2005536355 hasConceptScore W2005536355C40969351 @default.
- W2005536355 hasConceptScore W2005536355C41008148 @default.
- W2005536355 hasConceptScore W2005536355C74193536 @default.
- W2005536355 hasLocation W20055363551 @default.
- W2005536355 hasOpenAccess W2005536355 @default.
- W2005536355 hasPrimaryLocation W20055363551 @default.
- W2005536355 hasRelatedWork W1506229137 @default.
- W2005536355 hasRelatedWork W1996628468 @default.
- W2005536355 hasRelatedWork W2057906900 @default.
- W2005536355 hasRelatedWork W2065857332 @default.
- W2005536355 hasRelatedWork W2066384148 @default.
- W2005536355 hasRelatedWork W2107545280 @default.
- W2005536355 hasRelatedWork W2109007838 @default.
- W2005536355 hasRelatedWork W2113719240 @default.
- W2005536355 hasRelatedWork W2115208651 @default.
- W2005536355 hasRelatedWork W2158552994 @default.
- W2005536355 hasRelatedWork W2163307649 @default.
- W2005536355 hasRelatedWork W2366496188 @default.
- W2005536355 hasRelatedWork W2381146404 @default.
- W2005536355 hasRelatedWork W2804655408 @default.
- W2005536355 hasRelatedWork W2896704726 @default.
- W2005536355 hasRelatedWork W2899780697 @default.
- W2005536355 hasRelatedWork W2900187162 @default.
- W2005536355 hasRelatedWork W2942872985 @default.
- W2005536355 hasRelatedWork W3012052714 @default.
- W2005536355 hasRelatedWork W2146068149 @default.
- W2005536355 isParatext "false" @default.
- W2005536355 isRetracted "false" @default.
- W2005536355 magId "2005536355" @default.
- W2005536355 workType "article" @default.