FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3092357178> ?p ?o ?g. }

• W3092357178 abstract "Training Convolutional Neural Networks (CNNs) usually requires a large number of computational resources. In this paper, SparseTrain is proposed to accelerate CNN training by fully exploiting the sparsity. It mainly involves three levels of innovations: activation gradients pruning algorithm, sparse training dataflow, and accelerator architecture. By applying a stochastic pruning algorithm on each layer, the sparsity of back-propagation gradients can be increased dramatically without degrading training accuracy and convergence rate. Moreover, to utilize both natural sparsity (resulted from ReLU or Pooling layers) and artificial sparsity (brought by pruning algorithm), a sparse-aware architecture is proposed for training acceleration. This architecture supports forward and back-propagation of CNN by adopting 1-Dimensional convolution dataflow. We have built a cycle-accurate architecture simulator to evaluate the performance and efficiency based on the synthesized design with 14nm FinFET technologies. Evaluation results on AlexNet/ResNet show that SparseTrain could achieve about 2.7× speedup and 2.2× energy efficiency improvement on average compared with the original training process." @default.
• W3092357178 created "2020-10-15" @default.
• W3092357178 creator A5003076964 @default.
• W3092357178 creator A5036239104 @default.
• W3092357178 creator A5051712748 @default.
• W3092357178 creator A5053303853 @default.
• W3092357178 creator A5058073627 @default.
• W3092357178 creator A5066473925 @default.
• W3092357178 creator A5075136106 @default.
• W3092357178 creator A5082464125 @default.
• W3092357178 date "2020-07-01" @default.
• W3092357178 modified "2023-10-11" @default.
• W3092357178 title "SparseTrain: Exploiting Dataflow Sparsity for Efficient Convolutional Neural Networks Training" @default.
• W3092357178 cites W2041811089 @default.
• W3092357178 cites W2140609507 @default.
• W3092357178 cites W2163605009 @default.
• W3092357178 cites W2194775991 @default.
• W3092357178 cites W2285660444 @default.
• W3092357178 cites W2442974303 @default.
• W3092357178 cites W2470394683 @default.
• W3092357178 cites W2605135468 @default.
• W3092357178 cites W2606722458 @default.
• W3092357178 cites W2613718673 @default.
• W3092357178 cites W2754526845 @default.
• W3092357178 cites W2801748224 @default.
• W3092357178 cites W2807239565 @default.
• W3092357178 cites W2905104204 @default.
• W3092357178 cites W2919115771 @default.
• W3092357178 cites W2921118685 @default.
• W3092357178 cites W2949989598 @default.
• W3092357178 cites W2962818002 @default.
• W3092357178 cites W2964004663 @default.
• W3092357178 cites W2964299589 @default.
• W3092357178 cites W2982644126 @default.
• W3092357178 cites W3107472389 @default.
• W3092357178 doi "https://doi.org/10.1109/dac18072.2020.9218710" @default.
• W3092357178 hasPublicationYear "2020" @default.
• W3092357178 type Work @default.
• W3092357178 sameAs 3092357178 @default.
• W3092357178 citedByCount "15" @default.
• W3092357178 countsByYear W30923571782019 @default.
• W3092357178 countsByYear W30923571782021 @default.
• W3092357178 countsByYear W30923571782022 @default.
• W3092357178 countsByYear W30923571782023 @default.
• W3092357178 crossrefType "proceedings-article" @default.
• W3092357178 hasAuthorship W3092357178A5003076964 @default.
• W3092357178 hasAuthorship W3092357178A5036239104 @default.
• W3092357178 hasAuthorship W3092357178A5051712748 @default.
• W3092357178 hasAuthorship W3092357178A5053303853 @default.
• W3092357178 hasAuthorship W3092357178A5058073627 @default.
• W3092357178 hasAuthorship W3092357178A5066473925 @default.
• W3092357178 hasAuthorship W3092357178A5075136106 @default.
• W3092357178 hasAuthorship W3092357178A5082464125 @default.
• W3092357178 hasBestOaLocation W30923571782 @default.
• W3092357178 hasConcept C108010975 @default.
• W3092357178 hasConcept C111919701 @default.
• W3092357178 hasConcept C113775141 @default.
• W3092357178 hasConcept C11413529 @default.
• W3092357178 hasConcept C154945302 @default.
• W3092357178 hasConcept C155032097 @default.
• W3092357178 hasConcept C162324750 @default.
• W3092357178 hasConcept C173608175 @default.
• W3092357178 hasConcept C176727019 @default.
• W3092357178 hasConcept C206688291 @default.
• W3092357178 hasConcept C21547014 @default.
• W3092357178 hasConcept C2777303404 @default.
• W3092357178 hasConcept C2778915421 @default.
• W3092357178 hasConcept C41008148 @default.
• W3092357178 hasConcept C45347329 @default.
• W3092357178 hasConcept C50522688 @default.
• W3092357178 hasConcept C50644808 @default.
• W3092357178 hasConcept C6557445 @default.
• W3092357178 hasConcept C68339613 @default.
• W3092357178 hasConcept C70437156 @default.
• W3092357178 hasConcept C81363708 @default.
• W3092357178 hasConcept C86803240 @default.
• W3092357178 hasConcept C96324660 @default.
• W3092357178 hasConcept C98045186 @default.
• W3092357178 hasConceptScore W3092357178C108010975 @default.
• W3092357178 hasConceptScore W3092357178C111919701 @default.
• W3092357178 hasConceptScore W3092357178C113775141 @default.
• W3092357178 hasConceptScore W3092357178C11413529 @default.
• W3092357178 hasConceptScore W3092357178C154945302 @default.
• W3092357178 hasConceptScore W3092357178C155032097 @default.
• W3092357178 hasConceptScore W3092357178C162324750 @default.
• W3092357178 hasConceptScore W3092357178C173608175 @default.
• W3092357178 hasConceptScore W3092357178C176727019 @default.
• W3092357178 hasConceptScore W3092357178C206688291 @default.
• W3092357178 hasConceptScore W3092357178C21547014 @default.
• W3092357178 hasConceptScore W3092357178C2777303404 @default.
• W3092357178 hasConceptScore W3092357178C2778915421 @default.
• W3092357178 hasConceptScore W3092357178C41008148 @default.
• W3092357178 hasConceptScore W3092357178C45347329 @default.
• W3092357178 hasConceptScore W3092357178C50522688 @default.
• W3092357178 hasConceptScore W3092357178C50644808 @default.
• W3092357178 hasConceptScore W3092357178C6557445 @default.
• W3092357178 hasConceptScore W3092357178C68339613 @default.
• W3092357178 hasConceptScore W3092357178C70437156 @default.
• W3092357178 hasConceptScore W3092357178C81363708 @default.
• W3092357178 hasConceptScore W3092357178C86803240 @default.

• next