Matches in SemOpenAlex for { <https://semopenalex.org/work/W3034783496> ?p ?o ?g. }
Showing items 1 to 86 of
86
with 100 items per page.
- W3034783496 endingPage "969" @default.
- W3034783496 startingPage "960" @default.
- W3034783496 abstract "Deep neural networks are typically initialized with random weights, with variances chosen to facilitate signal propagation and stable gradients. It is also believed that diversity of features is an important property of these initializations. We construct a deep convolutional network with identical features by initializing almost all the weights to $0$. The architecture also enables perfect signal propagation and stable gradients, and achieves high accuracy on standard benchmarks. This indicates that random, diverse initializations are textit{not} necessary for training neural networks. An essential element in training this network is a mechanism of symmetry breaking; we study this phenomenon and find that standard GPU operations, which are non-deterministic, can serve as a sufficient source of symmetry breaking to enable training." @default.
- W3034783496 created "2020-06-19" @default.
- W3034783496 creator A5007253939 @default.
- W3034783496 creator A5017296839 @default.
- W3034783496 creator A5025123281 @default.
- W3034783496 date "2020-07-12" @default.
- W3034783496 modified "2023-09-28" @default.
- W3034783496 title "Beyond Signal Propagation: Is Feature Diversity Necessary in Deep Neural Network Initialization?" @default.
- W3034783496 hasPublicationYear "2020" @default.
- W3034783496 type Work @default.
- W3034783496 sameAs 3034783496 @default.
- W3034783496 citedByCount "3" @default.
- W3034783496 countsByYear W30347834962020 @default.
- W3034783496 countsByYear W30347834962021 @default.
- W3034783496 crossrefType "proceedings-article" @default.
- W3034783496 hasAuthorship W3034783496A5007253939 @default.
- W3034783496 hasAuthorship W3034783496A5017296839 @default.
- W3034783496 hasAuthorship W3034783496A5025123281 @default.
- W3034783496 hasConcept C108583219 @default.
- W3034783496 hasConcept C111472728 @default.
- W3034783496 hasConcept C11413529 @default.
- W3034783496 hasConcept C114466953 @default.
- W3034783496 hasConcept C138885662 @default.
- W3034783496 hasConcept C153180895 @default.
- W3034783496 hasConcept C154945302 @default.
- W3034783496 hasConcept C155032097 @default.
- W3034783496 hasConcept C189950617 @default.
- W3034783496 hasConcept C199360897 @default.
- W3034783496 hasConcept C2524010 @default.
- W3034783496 hasConcept C2776401178 @default.
- W3034783496 hasConcept C2779843651 @default.
- W3034783496 hasConcept C2779886137 @default.
- W3034783496 hasConcept C2780801425 @default.
- W3034783496 hasConcept C33923547 @default.
- W3034783496 hasConcept C41008148 @default.
- W3034783496 hasConcept C41895202 @default.
- W3034783496 hasConcept C50644808 @default.
- W3034783496 hasConcept C81363708 @default.
- W3034783496 hasConceptScore W3034783496C108583219 @default.
- W3034783496 hasConceptScore W3034783496C111472728 @default.
- W3034783496 hasConceptScore W3034783496C11413529 @default.
- W3034783496 hasConceptScore W3034783496C114466953 @default.
- W3034783496 hasConceptScore W3034783496C138885662 @default.
- W3034783496 hasConceptScore W3034783496C153180895 @default.
- W3034783496 hasConceptScore W3034783496C154945302 @default.
- W3034783496 hasConceptScore W3034783496C155032097 @default.
- W3034783496 hasConceptScore W3034783496C189950617 @default.
- W3034783496 hasConceptScore W3034783496C199360897 @default.
- W3034783496 hasConceptScore W3034783496C2524010 @default.
- W3034783496 hasConceptScore W3034783496C2776401178 @default.
- W3034783496 hasConceptScore W3034783496C2779843651 @default.
- W3034783496 hasConceptScore W3034783496C2779886137 @default.
- W3034783496 hasConceptScore W3034783496C2780801425 @default.
- W3034783496 hasConceptScore W3034783496C33923547 @default.
- W3034783496 hasConceptScore W3034783496C41008148 @default.
- W3034783496 hasConceptScore W3034783496C41895202 @default.
- W3034783496 hasConceptScore W3034783496C50644808 @default.
- W3034783496 hasConceptScore W3034783496C81363708 @default.
- W3034783496 hasOpenAccess W3034783496 @default.
- W3034783496 hasRelatedWork W1989422077 @default.
- W3034783496 hasRelatedWork W2113517874 @default.
- W3034783496 hasRelatedWork W2473118437 @default.
- W3034783496 hasRelatedWork W2593110912 @default.
- W3034783496 hasRelatedWork W2743693791 @default.
- W3034783496 hasRelatedWork W2762534372 @default.
- W3034783496 hasRelatedWork W2812009592 @default.
- W3034783496 hasRelatedWork W2905801789 @default.
- W3034783496 hasRelatedWork W2906264008 @default.
- W3034783496 hasRelatedWork W2945799378 @default.
- W3034783496 hasRelatedWork W2964658395 @default.
- W3034783496 hasRelatedWork W2987612595 @default.
- W3034783496 hasRelatedWork W2996243516 @default.
- W3034783496 hasRelatedWork W3003481009 @default.
- W3034783496 hasRelatedWork W3021623424 @default.
- W3034783496 hasRelatedWork W3040425217 @default.
- W3034783496 hasRelatedWork W3104371525 @default.
- W3034783496 hasRelatedWork W3174185214 @default.
- W3034783496 hasRelatedWork W3183466187 @default.
- W3034783496 hasRelatedWork W3201496309 @default.
- W3034783496 hasVolume "1" @default.
- W3034783496 isParatext "false" @default.
- W3034783496 isRetracted "false" @default.
- W3034783496 magId "3034783496" @default.
- W3034783496 workType "article" @default.