Matches in SemOpenAlex for { <https://semopenalex.org/work/W3204165788> ?p ?o ?g. }
Showing items 1 to 90 of
90
with 100 items per page.
- W3204165788 abstract "Deep neural networks (DNNs) sustain high performance in today’s data processing applications. DNN inference is resource-intensive thus is difficult to fit into a mobile device. To solve this problem, recent advances unleash DNN services using the edge computing paradigm. The existing approaches split a DNN into two parts and deploy the two partitions to computation nodes at two edge computing tiers. Nonetheless, these methods overlook collaborative device-edge-cloud computation resources. Besides, previous algorithms demand the whole DNN re-partitioning to adapt to computation resource changes and network dynamics. Moreover, for resource-demanding convolutional layers, prior works do not give a parallel processing strategy without loss of accuracy at the edge side. To tackle these issues, we propose D <sup xmlns:mml=http://www.w3.org/1998/Math/MathML xmlns:xlink=http://www.w3.org/1999/xlink>3</sup> , a dynamic DNN decomposition system for synergistic inference without precision loss. The proposed system introduces a heuristic algorithm named horizontal partition algorithm to split a DNN into three parts. The algorithm partially adjust the partitions at run time according to processing time and network conditions. At the edge side, a vertical separation module separates feature maps into tiles that can be independently run on different edge nodes in parallel. Extensive quantitative evaluation of five popular DNNs illustrates that D <sup xmlns:mml=http://www.w3.org/1998/Math/MathML xmlns:xlink=http://www.w3.org/1999/xlink>3</sup> outperforms the state-of-the-art counterparts up to 3.4× in end-to-end DNN inference time and reduces backbone network communication overhead up to 3.68×." @default.
- W3204165788 created "2021-10-11" @default.
- W3204165788 creator A5002358967 @default.
- W3204165788 creator A5039038654 @default.
- W3204165788 creator A5039356472 @default.
- W3204165788 creator A5076403451 @default.
- W3204165788 creator A5082387283 @default.
- W3204165788 creator A5085217070 @default.
- W3204165788 date "2021-07-01" @default.
- W3204165788 modified "2023-10-16" @default.
- W3204165788 title "Dynamic DNN Decomposition for Lossless Synergistic Inference" @default.
- W3204165788 cites W2061344791 @default.
- W3204165788 cites W2108598243 @default.
- W3204165788 cites W2194775991 @default.
- W3204165788 cites W2605258629 @default.
- W3204165788 cites W2612193523 @default.
- W3204165788 cites W2892952080 @default.
- W3204165788 cites W2896180420 @default.
- W3204165788 cites W2920031528 @default.
- W3204165788 cites W2921456883 @default.
- W3204165788 cites W2950865323 @default.
- W3204165788 cites W2962814013 @default.
- W3204165788 cites W2962883027 @default.
- W3204165788 cites W2983440318 @default.
- W3204165788 cites W3047565185 @default.
- W3204165788 doi "https://doi.org/10.1109/icdcsw53096.2021.00010" @default.
- W3204165788 hasPublicationYear "2021" @default.
- W3204165788 type Work @default.
- W3204165788 sameAs 3204165788 @default.
- W3204165788 citedByCount "5" @default.
- W3204165788 countsByYear W32041657882022 @default.
- W3204165788 countsByYear W32041657882023 @default.
- W3204165788 crossrefType "proceedings-article" @default.
- W3204165788 hasAuthorship W3204165788A5002358967 @default.
- W3204165788 hasAuthorship W3204165788A5039038654 @default.
- W3204165788 hasAuthorship W3204165788A5039356472 @default.
- W3204165788 hasAuthorship W3204165788A5076403451 @default.
- W3204165788 hasAuthorship W3204165788A5082387283 @default.
- W3204165788 hasAuthorship W3204165788A5085217070 @default.
- W3204165788 hasBestOaLocation W32041657882 @default.
- W3204165788 hasConcept C111919701 @default.
- W3204165788 hasConcept C11413529 @default.
- W3204165788 hasConcept C114614502 @default.
- W3204165788 hasConcept C120314980 @default.
- W3204165788 hasConcept C138236772 @default.
- W3204165788 hasConcept C154945302 @default.
- W3204165788 hasConcept C162307627 @default.
- W3204165788 hasConcept C173801870 @default.
- W3204165788 hasConcept C2776214188 @default.
- W3204165788 hasConcept C2778456923 @default.
- W3204165788 hasConcept C33923547 @default.
- W3204165788 hasConcept C41008148 @default.
- W3204165788 hasConcept C42812 @default.
- W3204165788 hasConcept C45374587 @default.
- W3204165788 hasConcept C79974875 @default.
- W3204165788 hasConcept C80444323 @default.
- W3204165788 hasConceptScore W3204165788C111919701 @default.
- W3204165788 hasConceptScore W3204165788C11413529 @default.
- W3204165788 hasConceptScore W3204165788C114614502 @default.
- W3204165788 hasConceptScore W3204165788C120314980 @default.
- W3204165788 hasConceptScore W3204165788C138236772 @default.
- W3204165788 hasConceptScore W3204165788C154945302 @default.
- W3204165788 hasConceptScore W3204165788C162307627 @default.
- W3204165788 hasConceptScore W3204165788C173801870 @default.
- W3204165788 hasConceptScore W3204165788C2776214188 @default.
- W3204165788 hasConceptScore W3204165788C2778456923 @default.
- W3204165788 hasConceptScore W3204165788C33923547 @default.
- W3204165788 hasConceptScore W3204165788C41008148 @default.
- W3204165788 hasConceptScore W3204165788C42812 @default.
- W3204165788 hasConceptScore W3204165788C45374587 @default.
- W3204165788 hasConceptScore W3204165788C79974875 @default.
- W3204165788 hasConceptScore W3204165788C80444323 @default.
- W3204165788 hasLocation W32041657881 @default.
- W3204165788 hasLocation W32041657882 @default.
- W3204165788 hasOpenAccess W3204165788 @default.
- W3204165788 hasPrimaryLocation W32041657881 @default.
- W3204165788 hasRelatedWork W2534668683 @default.
- W3204165788 hasRelatedWork W2942586735 @default.
- W3204165788 hasRelatedWork W3009018976 @default.
- W3204165788 hasRelatedWork W3013760193 @default.
- W3204165788 hasRelatedWork W3211931762 @default.
- W3204165788 hasRelatedWork W4225757241 @default.
- W3204165788 hasRelatedWork W4307482744 @default.
- W3204165788 hasRelatedWork W4375928818 @default.
- W3204165788 hasRelatedWork W4385414328 @default.
- W3204165788 hasRelatedWork W4385586765 @default.
- W3204165788 isParatext "false" @default.
- W3204165788 isRetracted "false" @default.
- W3204165788 magId "3204165788" @default.
- W3204165788 workType "article" @default.