FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W4221058856> ?p ?o ?g. }

• W4221058856 endingPage "29" @default.
• W4221058856 startingPage "1" @default.
• W4221058856 abstract "Machine learning (ML) on resource-constrained edge devices is expensive and often requires offloading computation to the cloud, which may compromise the privacy of user data. In contrast, the type of data processed at edge devices is user-specific and limited to a few inference classes. In this work, we explore building smaller, user-specific machine learning models, rather than utilizing a generic, compute-intensive machine learning model that caters to a diverse range of users. We first present a hardware-friendly, lightweight pruning technique to create user-specific models directly on mobile platforms, while simultaneously executing inferences. The proposed technique leverages compute sharing between pruning and inference, customizes the backward pass of training, and chooses a pruning granularity for efficient processing on edge. We then propose architectural support to prune user-specific models on a systolic edge ML inference accelerator. We demonstrate that user-specific models provide a speedup of 2.9× and 2.3× on the mobile CPUs for the ResNet-50 and Inception-V3 models." @default.
• W4221058856 created "2022-04-03" @default.
• W4221058856 creator A5027544576 @default.
• W4221058856 creator A5030335506 @default.
• W4221058856 creator A5069137406 @default.
• W4221058856 date "2022-09-30" @default.
• W4221058856 modified "2023-10-16" @default.
• W4221058856 title "Hardware-friendly User-specific Machine Learning for Edge Devices" @default.
• W4221058856 cites W2097117768 @default.
• W4221058856 cites W2108598243 @default.
• W4221058856 cites W2183341477 @default.
• W4221058856 cites W2585720638 @default.
• W4221058856 cites W2604319603 @default.
• W4221058856 cites W2606722458 @default.
• W4221058856 cites W2725615981 @default.
• W4221058856 cites W2767101511 @default.
• W4221058856 cites W2791175987 @default.
• W4221058856 cites W2794478957 @default.
• W4221058856 cites W2810482148 @default.
• W4221058856 cites W2883283076 @default.
• W4221058856 cites W2883899390 @default.
• W4221058856 cites W2886851211 @default.
• W4221058856 cites W2906043559 @default.
• W4221058856 cites W2922395136 @default.
• W4221058856 cites W2931743911 @default.
• W4221058856 cites W2932154853 @default.
• W4221058856 cites W2934853022 @default.
• W4221058856 cites W2963145730 @default.
• W4221058856 cites W2963190722 @default.
• W4221058856 cites W2963363373 @default.
• W4221058856 cites W2964233199 @default.
• W4221058856 cites W3008026851 @default.
• W4221058856 cites W3008591352 @default.
• W4221058856 cites W3019721760 @default.
• W4221058856 cites W3031023929 @default.
• W4221058856 cites W3035377608 @default.
• W4221058856 cites W3090912412 @default.
• W4221058856 cites W3099365437 @default.
• W4221058856 cites W3100023286 @default.
• W4221058856 cites W3102587717 @default.
• W4221058856 cites W3102659484 @default.
• W4221058856 cites W3104849992 @default.
• W4221058856 cites W4235435541 @default.
• W4221058856 cites W4236868170 @default.
• W4221058856 cites W4244330903 @default.
• W4221058856 cites W4288079690 @default.
• W4221058856 doi "https://doi.org/10.1145/3524125" @default.
• W4221058856 hasPublicationYear "2022" @default.
• W4221058856 type Work @default.
• W4221058856 citedByCount "0" @default.
• W4221058856 crossrefType "journal-article" @default.
• W4221058856 hasAuthorship W4221058856A5027544576 @default.
• W4221058856 hasAuthorship W4221058856A5030335506 @default.
• W4221058856 hasAuthorship W4221058856A5069137406 @default.
• W4221058856 hasConcept C108010975 @default.
• W4221058856 hasConcept C111919701 @default.
• W4221058856 hasConcept C113775141 @default.
• W4221058856 hasConcept C119857082 @default.
• W4221058856 hasConcept C138236772 @default.
• W4221058856 hasConcept C154945302 @default.
• W4221058856 hasConcept C162307627 @default.
• W4221058856 hasConcept C173608175 @default.
• W4221058856 hasConcept C177774035 @default.
• W4221058856 hasConcept C186967261 @default.
• W4221058856 hasConcept C2776214188 @default.
• W4221058856 hasConcept C2778456923 @default.
• W4221058856 hasConcept C41008148 @default.
• W4221058856 hasConcept C61437842 @default.
• W4221058856 hasConcept C6557445 @default.
• W4221058856 hasConcept C68339613 @default.
• W4221058856 hasConcept C79974875 @default.
• W4221058856 hasConcept C86803240 @default.
• W4221058856 hasConceptScore W4221058856C108010975 @default.
• W4221058856 hasConceptScore W4221058856C111919701 @default.
• W4221058856 hasConceptScore W4221058856C113775141 @default.
• W4221058856 hasConceptScore W4221058856C119857082 @default.
• W4221058856 hasConceptScore W4221058856C138236772 @default.
• W4221058856 hasConceptScore W4221058856C154945302 @default.
• W4221058856 hasConceptScore W4221058856C162307627 @default.
• W4221058856 hasConceptScore W4221058856C173608175 @default.
• W4221058856 hasConceptScore W4221058856C177774035 @default.
• W4221058856 hasConceptScore W4221058856C186967261 @default.
• W4221058856 hasConceptScore W4221058856C2776214188 @default.
• W4221058856 hasConceptScore W4221058856C2778456923 @default.
• W4221058856 hasConceptScore W4221058856C41008148 @default.
• W4221058856 hasConceptScore W4221058856C61437842 @default.
• W4221058856 hasConceptScore W4221058856C6557445 @default.
• W4221058856 hasConceptScore W4221058856C68339613 @default.
• W4221058856 hasConceptScore W4221058856C79974875 @default.
• W4221058856 hasConceptScore W4221058856C86803240 @default.
• W4221058856 hasFunder F4320306087 @default.
• W4221058856 hasIssue "5" @default.
• W4221058856 hasLocation W42210588561 @default.
• W4221058856 hasOpenAccess W4221058856 @default.
• W4221058856 hasPrimaryLocation W42210588561 @default.
• W4221058856 hasRelatedWork W3098411449 @default.
• W4221058856 hasRelatedWork W3177201542 @default.
• W4221058856 hasRelatedWork W3211489647 @default.

• next