Matches in SemOpenAlex for { <https://semopenalex.org/work/W4385486938> ?p ?o ?g. }
Showing items 1 to 97 of
97
with 100 items per page.
- W4385486938 endingPage "444" @default.
- W4385486938 startingPage "432" @default.
- W4385486938 abstract "Deep Neural Network (DNN) models are usually trained sequentially from one layer to another, which causes forward, backward and update locking problems, leading to poor performance in terms of training time. The existing parallel strategies to mitigate these problems provide suboptimal runtime performance. In this work, we have proposed a novel layer-wise partitioning and merging, forward and backward pass parallel framework to provide better training performance. The novelty of the proposed work consists of 1) a layer-wise partition and merging model which can minimise communication overhead between devices without the memory cost of existing strategies during the training process; 2) a forward pass and backward pass parallelisation to address the update locking problem and minimise the total training cost. The experimental evaluation on real use cases shows that the proposed method outperforms the state-of-the-art approaches in terms of training speed; and achieves almost linear speedup without compromising the accuracy performance of the non-parallel approach." @default.
- W4385486938 created "2023-08-03" @default.
- W4385486938 creator A5023625466 @default.
- W4385486938 creator A5031804038 @default.
- W4385486938 creator A5044191256 @default.
- W4385486938 creator A5069957801 @default.
- W4385486938 creator A5082117229 @default.
- W4385486938 creator A5089363536 @default.
- W4385486938 creator A5090352566 @default.
- W4385486938 date "2023-12-01" @default.
- W4385486938 modified "2023-10-17" @default.
- W4385486938 title "Layer-wise partitioning and merging for efficient and scalable deep learning" @default.
- W4385486938 cites W2183341477 @default.
- W4385486938 cites W2185864411 @default.
- W4385486938 cites W2752782242 @default.
- W4385486938 cites W2951341874 @default.
- W4385486938 cites W2964149935 @default.
- W4385486938 cites W2964223234 @default.
- W4385486938 cites W2991615366 @default.
- W4385486938 cites W3012468284 @default.
- W4385486938 cites W3046253174 @default.
- W4385486938 cites W3129469040 @default.
- W4385486938 cites W3167669755 @default.
- W4385486938 doi "https://doi.org/10.1016/j.future.2023.07.043" @default.
- W4385486938 hasPublicationYear "2023" @default.
- W4385486938 type Work @default.
- W4385486938 citedByCount "0" @default.
- W4385486938 crossrefType "journal-article" @default.
- W4385486938 hasAuthorship W4385486938A5023625466 @default.
- W4385486938 hasAuthorship W4385486938A5031804038 @default.
- W4385486938 hasAuthorship W4385486938A5044191256 @default.
- W4385486938 hasAuthorship W4385486938A5069957801 @default.
- W4385486938 hasAuthorship W4385486938A5082117229 @default.
- W4385486938 hasAuthorship W4385486938A5089363536 @default.
- W4385486938 hasAuthorship W4385486938A5090352566 @default.
- W4385486938 hasBestOaLocation W43854869381 @default.
- W4385486938 hasConcept C111919701 @default.
- W4385486938 hasConcept C113775141 @default.
- W4385486938 hasConcept C114614502 @default.
- W4385486938 hasConcept C120314980 @default.
- W4385486938 hasConcept C138885662 @default.
- W4385486938 hasConcept C154945302 @default.
- W4385486938 hasConcept C173608175 @default.
- W4385486938 hasConcept C178790620 @default.
- W4385486938 hasConcept C185592680 @default.
- W4385486938 hasConcept C27206212 @default.
- W4385486938 hasConcept C2778738651 @default.
- W4385486938 hasConcept C2779227376 @default.
- W4385486938 hasConcept C2779960059 @default.
- W4385486938 hasConcept C33923547 @default.
- W4385486938 hasConcept C41008148 @default.
- W4385486938 hasConcept C42812 @default.
- W4385486938 hasConcept C48044578 @default.
- W4385486938 hasConcept C68339613 @default.
- W4385486938 hasConcept C77088390 @default.
- W4385486938 hasConcept C98045186 @default.
- W4385486938 hasConceptScore W4385486938C111919701 @default.
- W4385486938 hasConceptScore W4385486938C113775141 @default.
- W4385486938 hasConceptScore W4385486938C114614502 @default.
- W4385486938 hasConceptScore W4385486938C120314980 @default.
- W4385486938 hasConceptScore W4385486938C138885662 @default.
- W4385486938 hasConceptScore W4385486938C154945302 @default.
- W4385486938 hasConceptScore W4385486938C173608175 @default.
- W4385486938 hasConceptScore W4385486938C178790620 @default.
- W4385486938 hasConceptScore W4385486938C185592680 @default.
- W4385486938 hasConceptScore W4385486938C27206212 @default.
- W4385486938 hasConceptScore W4385486938C2778738651 @default.
- W4385486938 hasConceptScore W4385486938C2779227376 @default.
- W4385486938 hasConceptScore W4385486938C2779960059 @default.
- W4385486938 hasConceptScore W4385486938C33923547 @default.
- W4385486938 hasConceptScore W4385486938C41008148 @default.
- W4385486938 hasConceptScore W4385486938C42812 @default.
- W4385486938 hasConceptScore W4385486938C48044578 @default.
- W4385486938 hasConceptScore W4385486938C68339613 @default.
- W4385486938 hasConceptScore W4385486938C77088390 @default.
- W4385486938 hasConceptScore W4385486938C98045186 @default.
- W4385486938 hasLocation W43854869381 @default.
- W4385486938 hasLocation W43854869382 @default.
- W4385486938 hasLocation W43854869383 @default.
- W4385486938 hasOpenAccess W4385486938 @default.
- W4385486938 hasPrimaryLocation W43854869381 @default.
- W4385486938 hasRelatedWork W2013643406 @default.
- W4385486938 hasRelatedWork W2027972911 @default.
- W4385486938 hasRelatedWork W2058965144 @default.
- W4385486938 hasRelatedWork W2111089054 @default.
- W4385486938 hasRelatedWork W2115380918 @default.
- W4385486938 hasRelatedWork W2146343568 @default.
- W4385486938 hasRelatedWork W2150291671 @default.
- W4385486938 hasRelatedWork W2157978810 @default.
- W4385486938 hasRelatedWork W2164382479 @default.
- W4385486938 hasRelatedWork W98480971 @default.
- W4385486938 hasVolume "149" @default.
- W4385486938 isParatext "false" @default.
- W4385486938 isRetracted "false" @default.
- W4385486938 workType "article" @default.