DC Field | Value | Language |
---|---|---|
dc.contributor.author | Jung, Youngbeom | ko |
dc.contributor.author | Kim, Hyeonuk | ko |
dc.contributor.author | Choi, Yeongjae | ko |
dc.contributor.author | Kim, Lee-Sup | ko |
dc.date.accessioned | 2022-02-14T06:41:20Z | - |
dc.date.available | 2022-02-14T06:41:20Z | - |
dc.date.created | 2021-11-25 | - |
dc.date.created | 2021-11-25 | - |
dc.date.created | 2021-11-25 | - |
dc.date.issued | 2022-02 | - |
dc.identifier.citation | IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS II-EXPRESS BRIEFS, v.69, no.2, pp.609 - 613 | - |
dc.identifier.issn | 1549-7747 | - |
dc.identifier.uri | http://hdl.handle.net/10203/292219 | - |
dc.description.abstract | Quantization with low precision has become an essential technique for adopting deep neural networks in energy-and memory-constrained devices. However, there is a limit to the reducing precision by the inevitable loss of accuracy due to the quantization error. To overcome this obstacle, we propose methods reforming and quantizing a network that achieves high accuracy even at low precision without any runtime overhead in embedded accelerators. Our proposition consists of two analytical approaches: 1) network optimization to find the most error-resilient equivalent network in the precision-constrained environment and 2) quantization exploiting adaptive rounding offset control. The experimental results show accuracies of up to 98.31% and 99.96% of floating-point results in 6-bit and 8-bit quantization networks, respectively. Besides, our methods allow the lower precision accelerator design, reducing the energy consumption by 8.5%. | - |
dc.language | English | - |
dc.publisher | IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC | - |
dc.title | Quantization-Error-Robust Deep Neural Network for Embedded Accelerators | - |
dc.type | Article | - |
dc.identifier.wosid | 000748372000074 | - |
dc.identifier.scopusid | 2-s2.0-85124311225 | - |
dc.type.rims | ART | - |
dc.citation.volume | 69 | - |
dc.citation.issue | 2 | - |
dc.citation.beginningpage | 609 | - |
dc.citation.endingpage | 613 | - |
dc.citation.publicationname | IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS II-EXPRESS BRIEFS | - |
dc.identifier.doi | 10.1109/TCSII.2021.3103192 | - |
dc.contributor.localauthor | Kim, Lee-Sup | - |
dc.description.isOpenAccess | N | - |
dc.type.journalArticle | Article | - |
dc.subject.keywordAuthor | Quantization (signal)Adaptive systemsTrainingSignal resolutionNeural networksDeep learningCircuits and systemsDeep neural networkacceleratorquantizationrescaling equivalentadaptive rounding offset control | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.