FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2967460974> ?p ?o ?g. }

• W2967460974 abstract "Training deep convolutional neural networks such as VGG and ResNet by gradient descent is an expensive exercise requiring specialized hardware such as GPUs. Recent works have examined the possibility of approximating the gradient computation while maintaining the same convergence properties. While promising, the approximations only work on relatively small datasets such as MNIST. They also fail to achieve real wall-clock speedups due to lack of efficient GPU implementations of the proposed approximation methods. In this work, we explore three alternative methods to approximate gradients, with an efficient GPU kernel implementation for one of them. We achieve wall-clock speedup with ResNet-20 and VGG-19 on the CIFAR-10 dataset upwards of 7%, with a minimal loss in validation accuracy." @default.
• W2967460974 created "2019-08-22" @default.
• W2967460974 creator A5057879516 @default.
• W2967460974 creator A5085857010 @default.
• W2967460974 date "2019-08-15" @default.
• W2967460974 modified "2023-09-25" @default.
• W2967460974 title "Accelerated CNN Training Through Gradient Approximation" @default.
• W2967460974 cites W1667652561 @default.
• W2967460974 cites W1686810756 @default.
• W2967460974 cites W2194775991 @default.
• W2967460974 cites W2402144811 @default.
• W2967460974 cites W2552737632 @default.
• W2967460974 cites W2606722458 @default.
• W2967460974 cites W2617247391 @default.
• W2967460974 cites W2624273542 @default.
• W2967460974 cites W2754581707 @default.
• W2967460974 cites W2763421725 @default.
• W2967460974 cites W2769802733 @default.
• W2967460974 cites W2769856846 @default.
• W2967460974 cites W2785856116 @default.
• W2967460974 cites W2786738752 @default.
• W2967460974 cites W2789177660 @default.
• W2967460974 cites W2803140585 @default.
• W2967460974 cites W2806254827 @default.
• W2967460974 cites W2808794272 @default.
• W2967460974 cites W2884711234 @default.
• W2967460974 cites W2899171226 @default.
• W2967460974 cites W2908666743 @default.
• W2967460974 cites W2912283611 @default.
• W2967460974 cites W2934222362 @default.
• W2967460974 cites W2949117887 @default.
• W2967460974 cites W2953212265 @default.
• W2967460974 cites W2962747323 @default.
• W2967460974 cites W2963540381 @default.
• W2967460974 cites W2963616439 @default.
• W2967460974 cites W2963896595 @default.
• W2967460974 cites W2964004663 @default.
• W2967460974 cites W2964115671 @default.
• W2967460974 cites W2964299589 @default.
• W2967460974 cites W2990172850 @default.
• W2967460974 cites W3098508664 @default.
• W2967460974 cites W3101543398 @default.
• W2967460974 cites W3118608800 @default.
• W2967460974 doi "https://doi.org/10.48550/arxiv.1908.05460" @default.
• W2967460974 hasPublicationYear "2019" @default.
• W2967460974 type Work @default.
• W2967460974 sameAs 2967460974 @default.
• W2967460974 citedByCount "0" @default.
• W2967460974 crossrefType "posted-content" @default.
• W2967460974 hasAuthorship W2967460974A5057879516 @default.
• W2967460974 hasAuthorship W2967460974A5085857010 @default.
• W2967460974 hasBestOaLocation W29674609741 @default.
• W2967460974 hasConcept C108583219 @default.
• W2967460974 hasConcept C113775141 @default.
• W2967460974 hasConcept C11413529 @default.
• W2967460974 hasConcept C114614502 @default.
• W2967460974 hasConcept C153258448 @default.
• W2967460974 hasConcept C154945302 @default.
• W2967460974 hasConcept C162324750 @default.
• W2967460974 hasConcept C173608175 @default.
• W2967460974 hasConcept C190502265 @default.
• W2967460974 hasConcept C199360897 @default.
• W2967460974 hasConcept C206688291 @default.
• W2967460974 hasConcept C26713055 @default.
• W2967460974 hasConcept C2777303404 @default.
• W2967460974 hasConcept C2944601119 @default.
• W2967460974 hasConcept C33923547 @default.
• W2967460974 hasConcept C3826847 @default.
• W2967460974 hasConcept C41008148 @default.
• W2967460974 hasConcept C45374587 @default.
• W2967460974 hasConcept C459310 @default.
• W2967460974 hasConcept C50522688 @default.
• W2967460974 hasConcept C50644808 @default.
• W2967460974 hasConcept C68339613 @default.
• W2967460974 hasConcept C74193536 @default.
• W2967460974 hasConcept C81363708 @default.
• W2967460974 hasConceptScore W2967460974C108583219 @default.
• W2967460974 hasConceptScore W2967460974C113775141 @default.
• W2967460974 hasConceptScore W2967460974C11413529 @default.
• W2967460974 hasConceptScore W2967460974C114614502 @default.
• W2967460974 hasConceptScore W2967460974C153258448 @default.
• W2967460974 hasConceptScore W2967460974C154945302 @default.
• W2967460974 hasConceptScore W2967460974C162324750 @default.
• W2967460974 hasConceptScore W2967460974C173608175 @default.
• W2967460974 hasConceptScore W2967460974C190502265 @default.
• W2967460974 hasConceptScore W2967460974C199360897 @default.
• W2967460974 hasConceptScore W2967460974C206688291 @default.
• W2967460974 hasConceptScore W2967460974C26713055 @default.
• W2967460974 hasConceptScore W2967460974C2777303404 @default.
• W2967460974 hasConceptScore W2967460974C2944601119 @default.
• W2967460974 hasConceptScore W2967460974C33923547 @default.
• W2967460974 hasConceptScore W2967460974C3826847 @default.
• W2967460974 hasConceptScore W2967460974C41008148 @default.
• W2967460974 hasConceptScore W2967460974C45374587 @default.
• W2967460974 hasConceptScore W2967460974C459310 @default.
• W2967460974 hasConceptScore W2967460974C50522688 @default.
• W2967460974 hasConceptScore W2967460974C50644808 @default.
• W2967460974 hasConceptScore W2967460974C68339613 @default.
• W2967460974 hasConceptScore W2967460974C74193536 @default.
• W2967460974 hasConceptScore W2967460974C81363708 @default.

• next