FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W4312554056> ?p ?o ?g. }

• W4312554056 endingPage "16" @default.
• W4312554056 startingPage "4" @default.
• W4312554056 abstract "Approximate computing has emerged as a new paradigm that provides power-efficient and high-performance arithmetic designs by relaxing the stringent requirement of accuracy. Nonlinear functions (such as softmax, rectified linear unit (ReLU), Tanh, and Sigmoid) are extensively used in deep neural networks (DNNs). However, they incur significant power dissipation due to the high circuit complexity. As DNNs are error-tolerant, the design of approximation-linear functions is possible and desired. In this article, the design of an approximate softmax function (AxSF) is proposed. AxSF is based on a double hybrid structure (DHS). AxSF divides the input of the softmax function into two parts for different processing methods. The most significant bits (MSBs) are processed with lookup tables (LUTs) and an exact restoring array divider (EXDr). Taylor’s expansion and a logarithmic divider are used for the less significant bits (LSBs). An improved DHS (IDHS) is also proposed to reduce the hardware complexity. In IDHS, a novel Booth multiplier is utilized for the hybrid scheme to improve the partial product generation and compression, while the truncated implementation is applied to the divider unit. The proposed DHS and IDHS are compared with existing softmax designs. The results show that the proposed approximate softmax design reduces hardware by 48% and delay by 54% while retaining a high accuracy." @default.
• W4312554056 created "2023-01-05" @default.
• W4312554056 creator A5001979328 @default.
• W4312554056 creator A5004875361 @default.
• W4312554056 creator A5023734605 @default.
• W4312554056 creator A5056503588 @default.
• W4312554056 creator A5063931122 @default.
• W4312554056 date "2023-01-01" @default.
• W4312554056 modified "2023-10-14" @default.
• W4312554056 title "Approximate Softmax Functions for Energy-Efficient Deep Neural Networks" @default.
• W4312554056 cites W1498436455 @default.
• W4312554056 cites W1696859394 @default.
• W4312554056 cites W1978592082 @default.
• W4312554056 cites W1998824039 @default.
• W4312554056 cites W2016491712 @default.
• W4312554056 cites W2020217519 @default.
• W4312554056 cites W2026005150 @default.
• W4312554056 cites W2031489346 @default.
• W4312554056 cites W2045294186 @default.
• W4312554056 cites W2045881697 @default.
• W4312554056 cites W2064675550 @default.
• W4312554056 cites W2078078927 @default.
• W4312554056 cites W2084174584 @default.
• W4312554056 cites W2097117768 @default.
• W4312554056 cites W2105011602 @default.
• W4312554056 cites W2111013824 @default.
• W4312554056 cites W2121289914 @default.
• W4312554056 cites W2124616298 @default.
• W4312554056 cites W2136922672 @default.
• W4312554056 cites W2160815625 @default.
• W4312554056 cites W2161381512 @default.
• W4312554056 cites W2233797083 @default.
• W4312554056 cites W2338037132 @default.
• W4312554056 cites W2460228185 @default.
• W4312554056 cites W2585720638 @default.
• W4312554056 cites W2589986623 @default.
• W4312554056 cites W2606722458 @default.
• W4312554056 cites W2608093348 @default.
• W4312554056 cites W2613107232 @default.
• W4312554056 cites W2790285612 @default.
• W4312554056 cites W2795915628 @default.
• W4312554056 cites W2890165800 @default.
• W4312554056 cites W2903640996 @default.
• W4312554056 cites W2935524202 @default.
• W4312554056 cites W2937064326 @default.
• W4312554056 cites W2946480111 @default.
• W4312554056 cites W2962820060 @default.
• W4312554056 cites W2971601850 @default.
• W4312554056 cites W3014843990 @default.
• W4312554056 cites W3021972173 @default.
• W4312554056 cites W3036061420 @default.
• W4312554056 cites W3037801075 @default.
• W4312554056 cites W3046387293 @default.
• W4312554056 cites W3048414096 @default.
• W4312554056 cites W3054316493 @default.
• W4312554056 cites W3084604851 @default.
• W4312554056 cites W3085998684 @default.
• W4312554056 cites W3090555870 @default.
• W4312554056 cites W3120448834 @default.
• W4312554056 cites W3141007974 @default.
• W4312554056 cites W3142988090 @default.
• W4312554056 cites W3144137741 @default.
• W4312554056 cites W4234974086 @default.
• W4312554056 cites W639708223 @default.
• W4312554056 doi "https://doi.org/10.1109/tvlsi.2022.3224011" @default.
• W4312554056 hasPublicationYear "2023" @default.
• W4312554056 type Work @default.
• W4312554056 citedByCount "0" @default.
• W4312554056 crossrefType "journal-article" @default.
• W4312554056 hasAuthorship W4312554056A5001979328 @default.
• W4312554056 hasAuthorship W4312554056A5004875361 @default.
• W4312554056 hasAuthorship W4312554056A5023734605 @default.
• W4312554056 hasAuthorship W4312554056A5056503588 @default.
• W4312554056 hasAuthorship W4312554056A5063931122 @default.
• W4312554056 hasConcept C11413529 @default.
• W4312554056 hasConcept C134306372 @default.
• W4312554056 hasConcept C154945302 @default.
• W4312554056 hasConcept C188441871 @default.
• W4312554056 hasConcept C33923547 @default.
• W4312554056 hasConcept C39927690 @default.
• W4312554056 hasConcept C41008148 @default.
• W4312554056 hasConcept C50644808 @default.
• W4312554056 hasConcept C81388566 @default.
• W4312554056 hasConcept C92047909 @default.
• W4312554056 hasConcept C94375191 @default.
• W4312554056 hasConceptScore W4312554056C11413529 @default.
• W4312554056 hasConceptScore W4312554056C134306372 @default.
• W4312554056 hasConceptScore W4312554056C154945302 @default.
• W4312554056 hasConceptScore W4312554056C188441871 @default.
• W4312554056 hasConceptScore W4312554056C33923547 @default.
• W4312554056 hasConceptScore W4312554056C39927690 @default.
• W4312554056 hasConceptScore W4312554056C41008148 @default.
• W4312554056 hasConceptScore W4312554056C50644808 @default.
• W4312554056 hasConceptScore W4312554056C81388566 @default.
• W4312554056 hasConceptScore W4312554056C92047909 @default.
• W4312554056 hasConceptScore W4312554056C94375191 @default.
• W4312554056 hasFunder F4320321001 @default.
• W4312554056 hasIssue "1" @default.

• next