Matches in SemOpenAlex for { <https://semopenalex.org/work/W2810172011> ?p ?o ?g. }
- W2810172011 abstract "We introduce a principled approach for unsupervised structure learning of deep neural networks. We propose a new interpretation for depth and inter-layer connectivity where conditional independencies in the input distribution are encoded hierarchically in the network structure. Thus, the depth of the network is determined inherently. The proposed method casts the problem of neural network structure learning as a problem of Bayesian network structure learning. Then, instead of directly learning the discriminative structure, it learns a generative graph, constructs its stochastic inverse, and then constructs a discriminative graph. We prove that conditional-dependency relations among the latent variables in the generative graph are preserved in the class-conditional discriminative graph. We demonstrate on image classification benchmarks that the deepest layers (convolutional and dense) of common networks can be replaced by significantly smaller learned structures, while maintaining classification accuracy---state-of-the-art on tested benchmarks. Our structure learning algorithm requires a small computational cost and runs efficiently on a standard desktop CPU." @default.
- W2810172011 created "2018-07-10" @default.
- W2810172011 creator A5023367765 @default.
- W2810172011 creator A5042429340 @default.
- W2810172011 creator A5044277691 @default.
- W2810172011 creator A5056922604 @default.
- W2810172011 creator A5069937565 @default.
- W2810172011 date "2018-06-24" @default.
- W2810172011 modified "2023-09-25" @default.
- W2810172011 title "Constructing Deep Neural Networks by Bayesian Network Structure Learning" @default.
- W2810172011 cites W1686810756 @default.
- W2810172011 cites W1821462560 @default.
- W2810172011 cites W1935978687 @default.
- W2810172011 cites W2008906462 @default.
- W2810172011 cites W2083380015 @default.
- W2810172011 cites W2095705004 @default.
- W2810172011 cites W2097117768 @default.
- W2810172011 cites W2098694378 @default.
- W2810172011 cites W2102605133 @default.
- W2810172011 cites W2108598243 @default.
- W2810172011 cites W2112796928 @default.
- W2810172011 cites W2135181320 @default.
- W2810172011 cites W2136922672 @default.
- W2810172011 cites W2152332944 @default.
- W2810172011 cites W2152424459 @default.
- W2810172011 cites W2155541015 @default.
- W2810172011 cites W2158899491 @default.
- W2810172011 cites W2167215970 @default.
- W2810172011 cites W2178031510 @default.
- W2810172011 cites W2194775991 @default.
- W2810172011 cites W2253535400 @default.
- W2810172011 cites W2335728318 @default.
- W2810172011 cites W2401231614 @default.
- W2810172011 cites W2417988721 @default.
- W2810172011 cites W2463256602 @default.
- W2810172011 cites W2511730936 @default.
- W2810172011 cites W2534788641 @default.
- W2810172011 cites W2546302380 @default.
- W2810172011 cites W2553303224 @default.
- W2810172011 cites W2593744649 @default.
- W2810172011 cites W2771111398 @default.
- W2810172011 cites W2785253905 @default.
- W2810172011 cites W2949117887 @default.
- W2810172011 cites W2949264490 @default.
- W2810172011 cites W2951670162 @default.
- W2810172011 cites W2951886768 @default.
- W2810172011 cites W2952278564 @default.
- W2810172011 cites W2952838738 @default.
- W2810172011 cites W2963513744 @default.
- W2810172011 cites W2963674932 @default.
- W2810172011 cites W2964081807 @default.
- W2810172011 cites W2964121744 @default.
- W2810172011 cites W2964299589 @default.
- W2810172011 cites W2979006918 @default.
- W2810172011 cites W3118608800 @default.
- W2810172011 hasPublicationYear "2018" @default.
- W2810172011 type Work @default.
- W2810172011 sameAs 2810172011 @default.
- W2810172011 citedByCount "0" @default.
- W2810172011 crossrefType "posted-content" @default.
- W2810172011 hasAuthorship W2810172011A5023367765 @default.
- W2810172011 hasAuthorship W2810172011A5042429340 @default.
- W2810172011 hasAuthorship W2810172011A5044277691 @default.
- W2810172011 hasAuthorship W2810172011A5056922604 @default.
- W2810172011 hasAuthorship W2810172011A5069937565 @default.
- W2810172011 hasConcept C108583219 @default.
- W2810172011 hasConcept C119857082 @default.
- W2810172011 hasConcept C132525143 @default.
- W2810172011 hasConcept C153180895 @default.
- W2810172011 hasConcept C154945302 @default.
- W2810172011 hasConcept C167966045 @default.
- W2810172011 hasConcept C33724603 @default.
- W2810172011 hasConcept C39890363 @default.
- W2810172011 hasConcept C41008148 @default.
- W2810172011 hasConcept C50644808 @default.
- W2810172011 hasConcept C80444323 @default.
- W2810172011 hasConcept C81363708 @default.
- W2810172011 hasConcept C97931131 @default.
- W2810172011 hasConceptScore W2810172011C108583219 @default.
- W2810172011 hasConceptScore W2810172011C119857082 @default.
- W2810172011 hasConceptScore W2810172011C132525143 @default.
- W2810172011 hasConceptScore W2810172011C153180895 @default.
- W2810172011 hasConceptScore W2810172011C154945302 @default.
- W2810172011 hasConceptScore W2810172011C167966045 @default.
- W2810172011 hasConceptScore W2810172011C33724603 @default.
- W2810172011 hasConceptScore W2810172011C39890363 @default.
- W2810172011 hasConceptScore W2810172011C41008148 @default.
- W2810172011 hasConceptScore W2810172011C50644808 @default.
- W2810172011 hasConceptScore W2810172011C80444323 @default.
- W2810172011 hasConceptScore W2810172011C81363708 @default.
- W2810172011 hasConceptScore W2810172011C97931131 @default.
- W2810172011 hasLocation W28101720111 @default.
- W2810172011 hasOpenAccess W2810172011 @default.
- W2810172011 hasPrimaryLocation W28101720111 @default.
- W2810172011 hasRelatedWork W1992699959 @default.
- W2810172011 hasRelatedWork W2007477772 @default.
- W2810172011 hasRelatedWork W2029305120 @default.
- W2810172011 hasRelatedWork W2192598490 @default.
- W2810172011 hasRelatedWork W2299115575 @default.
- W2810172011 hasRelatedWork W2511871973 @default.