FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3036708415> ?p ?o ?g. }

• W3036708415 endingPage "3736" @default.
• W3036708415 startingPage "3726" @default.
• W3036708415 abstract "Convolutional layers are the core building blocks of Convolutional Neural Networks (CNNs). In this paper, we propose to augment a convolutional layer with an additional depthwise convolution, where each input channel is convolved with a different 2D kernel. The composition of the two convolutions constitutes an over-parameterization, since it adds learnable parameters, while the resulting linear operation can be expressed by a single convolution layer. We refer to this depthwise over-parameterized convolutional layer as DO-Conv, which is a novel way of over-parameterization. We show with extensive experiments that the mere replacement of conventional convolutional layers with DO-Conv layers boosts the performance of CNNs on many classical vision tasks, such as image classification, detection, and segmentation. Moreover, in the inference phase, the depthwise convolution is folded into the conventional convolution, reducing the computation to be exactly equivalent to that of a convolutional layer without over-parameterization. As DO-Conv introduces performance gains without incurring any computational complexity increase for inference, we advocate it as an alternative to the conventional convolutional layer. We open sourced an implementation of DO-Conv in Tensorflow, PyTorch and GluonCV at https://github.com/yangyanli/DO-Conv." @default.
• W3036708415 created "2020-06-25" @default.
• W3036708415 creator A5000322530 @default.
• W3036708415 creator A5010714340 @default.
• W3036708415 creator A5021053713 @default.
• W3036708415 creator A5032779538 @default.
• W3036708415 creator A5036688260 @default.
• W3036708415 creator A5040259500 @default.
• W3036708415 creator A5070514707 @default.
• W3036708415 creator A5087472282 @default.
• W3036708415 date "2022-01-01" @default.
• W3036708415 modified "2023-10-15" @default.
• W3036708415 title "DO-Conv: Depthwise Over-Parameterized Convolutional Layer" @default.
• W3036708415 cites W1677182931 @default.
• W3036708415 cites W2037227137 @default.
• W3036708415 cites W2097117768 @default.
• W3036708415 cites W2117539524 @default.
• W3036708415 cites W2194775991 @default.
• W3036708415 cites W2340897893 @default.
• W3036708415 cites W2531409750 @default.
• W3036708415 cites W2549139847 @default.
• W3036708415 cites W2601564443 @default.
• W3036708415 cites W2922509574 @default.
• W3036708415 cites W2963163009 @default.
• W3036708415 cites W2963351448 @default.
• W3036708415 cites W2963446712 @default.
• W3036708415 cites W2964137095 @default.
• W3036708415 cites W2966926453 @default.
• W3036708415 cites W2981609437 @default.
• W3036708415 cites W2982083293 @default.
• W3036708415 cites W2988396473 @default.
• W3036708415 cites W639708223 @default.
• W3036708415 doi "https://doi.org/10.1109/tip.2022.3175432" @default.
• W3036708415 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/35594231" @default.
• W3036708415 hasPublicationYear "2022" @default.
• W3036708415 type Work @default.
• W3036708415 sameAs 3036708415 @default.
• W3036708415 citedByCount "35" @default.
• W3036708415 countsByYear W30367084152021 @default.
• W3036708415 countsByYear W30367084152022 @default.
• W3036708415 countsByYear W30367084152023 @default.
• W3036708415 crossrefType "journal-article" @default.
• W3036708415 hasAuthorship W3036708415A5000322530 @default.
• W3036708415 hasAuthorship W3036708415A5010714340 @default.
• W3036708415 hasAuthorship W3036708415A5021053713 @default.
• W3036708415 hasAuthorship W3036708415A5032779538 @default.
• W3036708415 hasAuthorship W3036708415A5036688260 @default.
• W3036708415 hasAuthorship W3036708415A5040259500 @default.
• W3036708415 hasAuthorship W3036708415A5070514707 @default.
• W3036708415 hasAuthorship W3036708415A5087472282 @default.
• W3036708415 hasBestOaLocation W30367084152 @default.
• W3036708415 hasConcept C11413529 @default.
• W3036708415 hasConcept C114614502 @default.
• W3036708415 hasConcept C153180895 @default.
• W3036708415 hasConcept C154945302 @default.
• W3036708415 hasConcept C157899210 @default.
• W3036708415 hasConcept C165464430 @default.
• W3036708415 hasConcept C178790620 @default.
• W3036708415 hasConcept C185592680 @default.
• W3036708415 hasConcept C2776214188 @default.
• W3036708415 hasConcept C2779227376 @default.
• W3036708415 hasConcept C33923547 @default.
• W3036708415 hasConcept C41008148 @default.
• W3036708415 hasConcept C45347329 @default.
• W3036708415 hasConcept C45374587 @default.
• W3036708415 hasConcept C50644808 @default.
• W3036708415 hasConcept C57273362 @default.
• W3036708415 hasConcept C74193536 @default.
• W3036708415 hasConcept C81363708 @default.
• W3036708415 hasConceptScore W3036708415C11413529 @default.
• W3036708415 hasConceptScore W3036708415C114614502 @default.
• W3036708415 hasConceptScore W3036708415C153180895 @default.
• W3036708415 hasConceptScore W3036708415C154945302 @default.
• W3036708415 hasConceptScore W3036708415C157899210 @default.
• W3036708415 hasConceptScore W3036708415C165464430 @default.
• W3036708415 hasConceptScore W3036708415C178790620 @default.
• W3036708415 hasConceptScore W3036708415C185592680 @default.
• W3036708415 hasConceptScore W3036708415C2776214188 @default.
• W3036708415 hasConceptScore W3036708415C2779227376 @default.
• W3036708415 hasConceptScore W3036708415C33923547 @default.
• W3036708415 hasConceptScore W3036708415C41008148 @default.
• W3036708415 hasConceptScore W3036708415C45347329 @default.
• W3036708415 hasConceptScore W3036708415C45374587 @default.
• W3036708415 hasConceptScore W3036708415C50644808 @default.
• W3036708415 hasConceptScore W3036708415C57273362 @default.
• W3036708415 hasConceptScore W3036708415C74193536 @default.
• W3036708415 hasConceptScore W3036708415C81363708 @default.
• W3036708415 hasFunder F4320311026 @default.
• W3036708415 hasFunder F4320321001 @default.
• W3036708415 hasLocation W30367084151 @default.
• W3036708415 hasLocation W30367084152 @default.
• W3036708415 hasLocation W30367084153 @default.
• W3036708415 hasOpenAccess W3036708415 @default.
• W3036708415 hasPrimaryLocation W30367084151 @default.
• W3036708415 hasRelatedWork W2778739249 @default.
• W3036708415 hasRelatedWork W2796942851 @default.
• W3036708415 hasRelatedWork W2963993660 @default.
• W3036708415 hasRelatedWork W3005273712 @default.

• next