FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3216036760> ?p ?o ?g. }

• W3216036760 abstract "Recent advances in algorithm-hardware co-design for deep neural networks (DNNs) have demonstrated their potential in automatically designing neural architectures and hardware designs. Nevertheless, it is still a challenging optimization problem due to the expensive training cost and the time-consuming hardware implementation, which makes the exploration on the vast design space of neural architecture and hardware design intractable. In this paper, we demonstrate that our proposed approach is capable of locating designs on the Pareto frontier. This capability is enabled by a novel three-phase co-design framework, with the following new features: (a) decoupling DNN training from the design space exploration of hardware architecture and neural architecture, (b) providing a hardware-friendly neural architecture space by considering hardware characteristics in constructing the search cells, (c) adopting Gaussian process to predict accuracy, latency and power consumption to avoid time-consuming synthesis and place-and-route processes. In comparison with the manually-designed ResNet101, InceptionV2 and MobileNetV2, we can achieve up to 5% higher accuracy with up to <tex xmlns:mml=http://www.w3.org/1998/Math/MathML xmlns:xlink=http://www.w3.org/1999/xlink>$3times$</tex> speed up on the ImageNet dataset. Compared with other state-of-the-art co-design frameworks, our found network and hardware configuration can achieve 2% (~ 6% higher accuracy, <tex xmlns:mml=http://www.w3.org/1998/Math/MathML xmlns:xlink=http://www.w3.org/1999/xlink>$2timessim 26times$</tex> smaller latency and <tex xmlns:mml=http://www.w3.org/1998/Math/MathML xmlns:xlink=http://www.w3.org/1999/xlink>$8.5times$</tex> higher energy efficiency." @default.
• W3216036760 created "2021-12-06" @default.
• W3216036760 creator A5000359400 @default.
• W3216036760 creator A5054475218 @default.
• W3216036760 creator A5056354430 @default.
• W3216036760 creator A5057043409 @default.
• W3216036760 creator A5057940557 @default.
• W3216036760 creator A5059088556 @default.
• W3216036760 creator A5061397387 @default.
• W3216036760 date "2022-01-17" @default.
• W3216036760 modified "2023-09-30" @default.
• W3216036760 title "Algorithm and Hardware Co-design for Reconfigurable CNN Accelerator" @default.
• W3216036760 cites W2108598243 @default.
• W3216036760 cites W2183341477 @default.
• W3216036760 cites W2194775991 @default.
• W3216036760 cites W2906403229 @default.
• W3216036760 cites W2911884654 @default.
• W3216036760 cites W2951898149 @default.
• W3216036760 cites W2962856739 @default.
• W3216036760 cites W2963122961 @default.
• W3216036760 cites W2963163009 @default.
• W3216036760 cites W2967733054 @default.
• W3216036760 cites W2971769886 @default.
• W3216036760 cites W2981207549 @default.
• W3216036760 cites W2990728381 @default.
• W3216036760 cites W3013366513 @default.
• W3216036760 cites W3016142271 @default.
• W3216036760 cites W3036329912 @default.
• W3216036760 cites W3092020069 @default.
• W3216036760 cites W3092216346 @default.
• W3216036760 cites W3112721750 @default.
• W3216036760 cites W3114817143 @default.
• W3216036760 cites W3129147932 @default.
• W3216036760 cites W3168376295 @default.
• W3216036760 cites W3213229701 @default.
• W3216036760 doi "https://doi.org/10.1109/asp-dac52403.2022.9712541" @default.
• W3216036760 hasPublicationYear "2022" @default.
• W3216036760 type Work @default.
• W3216036760 sameAs 3216036760 @default.
• W3216036760 citedByCount "2" @default.
• W3216036760 countsByYear W32160367602023 @default.
• W3216036760 crossrefType "proceedings-article" @default.
• W3216036760 hasAuthorship W3216036760A5000359400 @default.
• W3216036760 hasAuthorship W3216036760A5054475218 @default.
• W3216036760 hasAuthorship W3216036760A5056354430 @default.
• W3216036760 hasAuthorship W3216036760A5057043409 @default.
• W3216036760 hasAuthorship W3216036760A5057940557 @default.
• W3216036760 hasAuthorship W3216036760A5059088556 @default.
• W3216036760 hasAuthorship W3216036760A5061397387 @default.
• W3216036760 hasBestOaLocation W32160367602 @default.
• W3216036760 hasConcept C113775141 @default.
• W3216036760 hasConcept C118524514 @default.
• W3216036760 hasConcept C123657996 @default.
• W3216036760 hasConcept C142362112 @default.
• W3216036760 hasConcept C149635348 @default.
• W3216036760 hasConcept C153349607 @default.
• W3216036760 hasConcept C154945302 @default.
• W3216036760 hasConcept C199360897 @default.
• W3216036760 hasConcept C2776221188 @default.
• W3216036760 hasConcept C2777904410 @default.
• W3216036760 hasConcept C41008148 @default.
• W3216036760 hasConcept C50644808 @default.
• W3216036760 hasConcept C65232700 @default.
• W3216036760 hasConcept C76155785 @default.
• W3216036760 hasConcept C82876162 @default.
• W3216036760 hasConcept C9390403 @default.
• W3216036760 hasConceptScore W3216036760C113775141 @default.
• W3216036760 hasConceptScore W3216036760C118524514 @default.
• W3216036760 hasConceptScore W3216036760C123657996 @default.
• W3216036760 hasConceptScore W3216036760C142362112 @default.
• W3216036760 hasConceptScore W3216036760C149635348 @default.
• W3216036760 hasConceptScore W3216036760C153349607 @default.
• W3216036760 hasConceptScore W3216036760C154945302 @default.
• W3216036760 hasConceptScore W3216036760C199360897 @default.
• W3216036760 hasConceptScore W3216036760C2776221188 @default.
• W3216036760 hasConceptScore W3216036760C2777904410 @default.
• W3216036760 hasConceptScore W3216036760C41008148 @default.
• W3216036760 hasConceptScore W3216036760C50644808 @default.
• W3216036760 hasConceptScore W3216036760C65232700 @default.
• W3216036760 hasConceptScore W3216036760C76155785 @default.
• W3216036760 hasConceptScore W3216036760C82876162 @default.
• W3216036760 hasConceptScore W3216036760C9390403 @default.
• W3216036760 hasFunder F4320321001 @default.
• W3216036760 hasFunder F4320322843 @default.
• W3216036760 hasLocation W32160367601 @default.
• W3216036760 hasLocation W32160367602 @default.
• W3216036760 hasLocation W32160367603 @default.
• W3216036760 hasLocation W32160367604 @default.
• W3216036760 hasOpenAccess W3216036760 @default.
• W3216036760 hasPrimaryLocation W32160367601 @default.
• W3216036760 hasRelatedWork W2020034302 @default.
• W3216036760 hasRelatedWork W2108386578 @default.
• W3216036760 hasRelatedWork W2147680600 @default.
• W3216036760 hasRelatedWork W2386834823 @default.
• W3216036760 hasRelatedWork W2761262947 @default.
• W3216036760 hasRelatedWork W3136178412 @default.
• W3216036760 hasRelatedWork W3177213278 @default.
• W3216036760 hasRelatedWork W3186087474 @default.
• W3216036760 hasRelatedWork W3188788977 @default.
• W3216036760 hasRelatedWork W4287273069 @default.

• next