Matches in SemOpenAlex for { <https://semopenalex.org/work/W4316658545> ?p ?o ?g. }
Showing items 1 to 90 of
90
with 100 items per page.
- W4316658545 abstract "Non-negative matrix factorization (NMF) is an ef-fective method for dimensionality reduction and sparse decom-position. This method has been of great interest to the scien-tific community in applications including signal processing, data mining, compression, and pattern recognition. However, NMF implies elevated computational costs in terms of performance and energy consumption, which is inadequate for embedded applications. To overcome this limitation, we implement the vector dot-product with hybrid logarithmic approximation as a hardware optimization approach. This technique accelerates floating-point computation, reduces energy consumption, and preserves accuracy. To demonstrate our approach, we employ a design exploration flow using high-level synthesis on an embedded FPGA. Compared with software solutions on ARM CPU, this hardware implementation accelerates the overall computation to decompose matrix by <tex xmlns:mml=http://www.w3.org/1998/Math/MathML xmlns:xlink=http://www.w3.org/1999/xlink>$5.597times$</tex> and reduces energy consumption by <tex xmlns:mml=http://www.w3.org/1998/Math/MathML xmlns:xlink=http://www.w3.org/1999/xlink>$69.323times$</tex> . Log approximation NMF combined with KNN(k-nearest neighbors) has only 2.38% decreasing accuracy compared with the result of KNN processing the matrix after floating-point NMF on MNIST. Further on, compared with a dedicated floating-point accelerator, the logarithmic approximation approach achieves <tex xmlns:mml=http://www.w3.org/1998/Math/MathML xmlns:xlink=http://www.w3.org/1999/xlink>$3.718times$</tex> acceleration and <tex xmlns:mml=http://www.w3.org/1998/Math/MathML xmlns:xlink=http://www.w3.org/1999/xlink>$8.345times$</tex> energy reduction. Compared with the fixed-point approach, our approach has an accuracy degradation of 1.93% on MNIST and an accuracy amelioration of 28.2% on the FASHION MNIST data set without pre-knowledge of the data range. Thus, our approach has better compatibility with the input data range." @default.
- W4316658545 created "2023-01-17" @default.
- W4316658545 creator A5007430313 @default.
- W4316658545 creator A5048315360 @default.
- W4316658545 creator A5050344559 @default.
- W4316658545 creator A5072973899 @default.
- W4316658545 date "2022-12-01" @default.
- W4316658545 modified "2023-09-30" @default.
- W4316658545 title "Accelerating Non-Negative Matrix Factorization on Embedded FPGA with Hybrid Logarithmic Dot-Product Approximation" @default.
- W4316658545 cites W1902027874 @default.
- W4316658545 cites W2020217519 @default.
- W4316658545 cites W2022978200 @default.
- W4316658545 cites W2030922317 @default.
- W4316658545 cites W2068142860 @default.
- W4316658545 cites W2112796928 @default.
- W4316658545 cites W2130855892 @default.
- W4316658545 cites W2136171036 @default.
- W4316658545 cites W2144920235 @default.
- W4316658545 cites W2163312093 @default.
- W4316658545 cites W2165685007 @default.
- W4316658545 cites W2731211745 @default.
- W4316658545 cites W2736045761 @default.
- W4316658545 cites W2794680152 @default.
- W4316658545 cites W2798956872 @default.
- W4316658545 cites W2804176448 @default.
- W4316658545 cites W2886937133 @default.
- W4316658545 cites W2900763943 @default.
- W4316658545 cites W3015461837 @default.
- W4316658545 cites W3100202987 @default.
- W4316658545 cites W3122703110 @default.
- W4316658545 cites W3171419617 @default.
- W4316658545 doi "https://doi.org/10.1109/mcsoc57363.2022.00070" @default.
- W4316658545 hasPublicationYear "2022" @default.
- W4316658545 type Work @default.
- W4316658545 citedByCount "0" @default.
- W4316658545 crossrefType "proceedings-article" @default.
- W4316658545 hasAuthorship W4316658545A5007430313 @default.
- W4316658545 hasAuthorship W4316658545A5048315360 @default.
- W4316658545 hasAuthorship W4316658545A5050344559 @default.
- W4316658545 hasAuthorship W4316658545A5072973899 @default.
- W4316658545 hasConcept C108583219 @default.
- W4316658545 hasConcept C11413529 @default.
- W4316658545 hasConcept C121332964 @default.
- W4316658545 hasConcept C13164978 @default.
- W4316658545 hasConcept C134306372 @default.
- W4316658545 hasConcept C152671427 @default.
- W4316658545 hasConcept C154945302 @default.
- W4316658545 hasConcept C158693339 @default.
- W4316658545 hasConcept C190502265 @default.
- W4316658545 hasConcept C33923547 @default.
- W4316658545 hasConcept C39927690 @default.
- W4316658545 hasConcept C41008148 @default.
- W4316658545 hasConcept C42355184 @default.
- W4316658545 hasConcept C42935608 @default.
- W4316658545 hasConcept C459310 @default.
- W4316658545 hasConcept C62520636 @default.
- W4316658545 hasConcept C9390403 @default.
- W4316658545 hasConceptScore W4316658545C108583219 @default.
- W4316658545 hasConceptScore W4316658545C11413529 @default.
- W4316658545 hasConceptScore W4316658545C121332964 @default.
- W4316658545 hasConceptScore W4316658545C13164978 @default.
- W4316658545 hasConceptScore W4316658545C134306372 @default.
- W4316658545 hasConceptScore W4316658545C152671427 @default.
- W4316658545 hasConceptScore W4316658545C154945302 @default.
- W4316658545 hasConceptScore W4316658545C158693339 @default.
- W4316658545 hasConceptScore W4316658545C190502265 @default.
- W4316658545 hasConceptScore W4316658545C33923547 @default.
- W4316658545 hasConceptScore W4316658545C39927690 @default.
- W4316658545 hasConceptScore W4316658545C41008148 @default.
- W4316658545 hasConceptScore W4316658545C42355184 @default.
- W4316658545 hasConceptScore W4316658545C42935608 @default.
- W4316658545 hasConceptScore W4316658545C459310 @default.
- W4316658545 hasConceptScore W4316658545C62520636 @default.
- W4316658545 hasConceptScore W4316658545C9390403 @default.
- W4316658545 hasLocation W43166585451 @default.
- W4316658545 hasOpenAccess W4316658545 @default.
- W4316658545 hasPrimaryLocation W43166585451 @default.
- W4316658545 hasRelatedWork W1095023703 @default.
- W4316658545 hasRelatedWork W2365743651 @default.
- W4316658545 hasRelatedWork W2805560392 @default.
- W4316658545 hasRelatedWork W2940225204 @default.
- W4316658545 hasRelatedWork W3108192785 @default.
- W4316658545 hasRelatedWork W3127488929 @default.
- W4316658545 hasRelatedWork W3205657855 @default.
- W4316658545 hasRelatedWork W4210526874 @default.
- W4316658545 hasRelatedWork W4226443758 @default.
- W4316658545 hasRelatedWork W4321786678 @default.
- W4316658545 isParatext "false" @default.
- W4316658545 isRetracted "false" @default.
- W4316658545 workType "article" @default.