Matches in SemOpenAlex for { <https://semopenalex.org/work/W2899206823> ?p ?o ?g. }
Showing items 1 to 82 of
82
with 100 items per page.
- W2899206823 endingPage "179" @default.
- W2899206823 startingPage "171" @default.
- W2899206823 abstract "The deep convolutional neural networks (CNN) have a vast amount of parameters, especially in the fully connected (FC) layers, which have become a bottleneck for real-time applications where processing latency is high due to computational cost. In this paper, we propose to optimize the FC layers in CNN via making it much slimmer. We make analysis of the statistical distribution of the weights in FC layer, and observe each column follows Gaussian distribution. Regression model analysis of the weights of FC layer based on Akaike information criteria and Bayesian information criterion demonstrates that they have Granger causality, which means the columns are correlated and they follow colored Gaussian distribution. Based on this distribution, we derive a CNN design and optimization theorem for FC layers from information theory point of view. The theorem provides two design criteria, rank and singular values. Further, we show that FC layer with weights of colored Gaussian is more efficient than that of white Gaussian. The optimization criteria is singular-values-based, so we apply singular value decomposition to find the maximal singular values and QR to identify the corresponding columns in FC layer. We evaluate our optimization approach to AlexNet and apply the slimmer CNN to ImageNet classification. Simulation results show our approach performs much better than random dropout. Specifically, with only around 28% of weights, the AlexNet could perform as well as the original AlexNet in terms of top one error and top five error." @default.
- W2899206823 created "2018-11-09" @default.
- W2899206823 creator A5090569244 @default.
- W2899206823 date "2020-04-01" @default.
- W2899206823 modified "2023-09-26" @default.
- W2899206823 title "Optimization for Deep Convolutional Neural Networks: How Slim Can It Go?" @default.
- W2899206823 cites W1590624540 @default.
- W2899206823 cites W1968157422 @default.
- W2899206823 cites W1968266099 @default.
- W2899206823 cites W1992633833 @default.
- W2899206823 cites W2097117768 @default.
- W2899206823 cites W2108598243 @default.
- W2899206823 cites W2116031726 @default.
- W2899206823 cites W2128728535 @default.
- W2899206823 cites W2194775991 @default.
- W2899206823 cites W2204679745 @default.
- W2899206823 cites W2614057673 @default.
- W2899206823 cites W2618530766 @default.
- W2899206823 cites W2743834698 @default.
- W2899206823 cites W2752782242 @default.
- W2899206823 cites W2758890981 @default.
- W2899206823 cites W2884552970 @default.
- W2899206823 cites W2885764039 @default.
- W2899206823 cites W2964184826 @default.
- W2899206823 cites W2979473749 @default.
- W2899206823 cites W4312258136 @default.
- W2899206823 doi "https://doi.org/10.1109/tetci.2018.2876573" @default.
- W2899206823 hasPublicationYear "2020" @default.
- W2899206823 type Work @default.
- W2899206823 sameAs 2899206823 @default.
- W2899206823 citedByCount "10" @default.
- W2899206823 countsByYear W28992068232020 @default.
- W2899206823 countsByYear W28992068232021 @default.
- W2899206823 countsByYear W28992068232022 @default.
- W2899206823 countsByYear W28992068232023 @default.
- W2899206823 crossrefType "journal-article" @default.
- W2899206823 hasAuthorship W2899206823A5090569244 @default.
- W2899206823 hasConcept C11413529 @default.
- W2899206823 hasConcept C119857082 @default.
- W2899206823 hasConcept C121332964 @default.
- W2899206823 hasConcept C126674687 @default.
- W2899206823 hasConcept C153180895 @default.
- W2899206823 hasConcept C154945302 @default.
- W2899206823 hasConcept C163716315 @default.
- W2899206823 hasConcept C22789450 @default.
- W2899206823 hasConcept C33923547 @default.
- W2899206823 hasConcept C41008148 @default.
- W2899206823 hasConcept C62520636 @default.
- W2899206823 hasConcept C81363708 @default.
- W2899206823 hasConceptScore W2899206823C11413529 @default.
- W2899206823 hasConceptScore W2899206823C119857082 @default.
- W2899206823 hasConceptScore W2899206823C121332964 @default.
- W2899206823 hasConceptScore W2899206823C126674687 @default.
- W2899206823 hasConceptScore W2899206823C153180895 @default.
- W2899206823 hasConceptScore W2899206823C154945302 @default.
- W2899206823 hasConceptScore W2899206823C163716315 @default.
- W2899206823 hasConceptScore W2899206823C22789450 @default.
- W2899206823 hasConceptScore W2899206823C33923547 @default.
- W2899206823 hasConceptScore W2899206823C41008148 @default.
- W2899206823 hasConceptScore W2899206823C62520636 @default.
- W2899206823 hasConceptScore W2899206823C81363708 @default.
- W2899206823 hasIssue "2" @default.
- W2899206823 hasLocation W28992068231 @default.
- W2899206823 hasOpenAccess W2899206823 @default.
- W2899206823 hasPrimaryLocation W28992068231 @default.
- W2899206823 hasRelatedWork W2068905082 @default.
- W2899206823 hasRelatedWork W2080597906 @default.
- W2899206823 hasRelatedWork W2748454020 @default.
- W2899206823 hasRelatedWork W2767651786 @default.
- W2899206823 hasRelatedWork W2912288872 @default.
- W2899206823 hasRelatedWork W3016958897 @default.
- W2899206823 hasRelatedWork W3181746755 @default.
- W2899206823 hasRelatedWork W4283379348 @default.
- W2899206823 hasRelatedWork W4312417841 @default.
- W2899206823 hasRelatedWork W564581980 @default.
- W2899206823 hasVolume "4" @default.
- W2899206823 isParatext "false" @default.
- W2899206823 isRetracted "false" @default.
- W2899206823 magId "2899206823" @default.
- W2899206823 workType "article" @default.