Matches in SemOpenAlex for { <https://semopenalex.org/work/W3105587016> ?p ?o ?g. }
Showing items 1 to 98 of
98
with 100 items per page.
- W3105587016 endingPage "4357" @default.
- W3105587016 startingPage "4346" @default.
- W3105587016 abstract "We challenge the longstanding assumption that the mean-field approximation for variational inference in Bayesian neural networks is severely restrictive, and show this is not the case in deep networks. We prove several results indicating that deep mean-field variational weight posteriors can induce similar distributions in function-space to those induced by shallower networks with complex weight posteriors. We validate our theoretical contributions empirically, both through examination of the weight posterior using Hamiltonian Monte Carlo in small models and by comparing diagonal- to structured-covariance in large settings. Since complex variational posteriors are often expensive and cumbersome to implement, our results suggest that using mean-field variational inference in a deeper model is both a practical and theoretically justified alternative to structured approximations." @default.
- W3105587016 created "2020-11-23" @default.
- W3105587016 creator A5014574161 @default.
- W3105587016 creator A5029186201 @default.
- W3105587016 creator A5051690066 @default.
- W3105587016 date "2020-02-10" @default.
- W3105587016 modified "2023-09-24" @default.
- W3105587016 title "Liberty or Depth: Deep Bayesian Neural Nets Do Not Need Complex Weight Posterior Approximations" @default.
- W3105587016 hasPublicationYear "2020" @default.
- W3105587016 type Work @default.
- W3105587016 sameAs 3105587016 @default.
- W3105587016 citedByCount "6" @default.
- W3105587016 countsByYear W31055870162020 @default.
- W3105587016 countsByYear W31055870162021 @default.
- W3105587016 crossrefType "proceedings-article" @default.
- W3105587016 hasAuthorship W3105587016A5014574161 @default.
- W3105587016 hasAuthorship W3105587016A5029186201 @default.
- W3105587016 hasAuthorship W3105587016A5051690066 @default.
- W3105587016 hasConcept C105795698 @default.
- W3105587016 hasConcept C107673813 @default.
- W3105587016 hasConcept C108583219 @default.
- W3105587016 hasConcept C111350023 @default.
- W3105587016 hasConcept C11413529 @default.
- W3105587016 hasConcept C121332964 @default.
- W3105587016 hasConcept C126255220 @default.
- W3105587016 hasConcept C130367717 @default.
- W3105587016 hasConcept C130787639 @default.
- W3105587016 hasConcept C13153151 @default.
- W3105587016 hasConcept C154945302 @default.
- W3105587016 hasConcept C160234255 @default.
- W3105587016 hasConcept C178650346 @default.
- W3105587016 hasConcept C19499675 @default.
- W3105587016 hasConcept C202213908 @default.
- W3105587016 hasConcept C202444582 @default.
- W3105587016 hasConcept C2524010 @default.
- W3105587016 hasConcept C2776214188 @default.
- W3105587016 hasConcept C28826006 @default.
- W3105587016 hasConcept C33923547 @default.
- W3105587016 hasConcept C41008148 @default.
- W3105587016 hasConcept C50644808 @default.
- W3105587016 hasConcept C57830394 @default.
- W3105587016 hasConcept C62520636 @default.
- W3105587016 hasConcept C9652623 @default.
- W3105587016 hasConceptScore W3105587016C105795698 @default.
- W3105587016 hasConceptScore W3105587016C107673813 @default.
- W3105587016 hasConceptScore W3105587016C108583219 @default.
- W3105587016 hasConceptScore W3105587016C111350023 @default.
- W3105587016 hasConceptScore W3105587016C11413529 @default.
- W3105587016 hasConceptScore W3105587016C121332964 @default.
- W3105587016 hasConceptScore W3105587016C126255220 @default.
- W3105587016 hasConceptScore W3105587016C130367717 @default.
- W3105587016 hasConceptScore W3105587016C130787639 @default.
- W3105587016 hasConceptScore W3105587016C13153151 @default.
- W3105587016 hasConceptScore W3105587016C154945302 @default.
- W3105587016 hasConceptScore W3105587016C160234255 @default.
- W3105587016 hasConceptScore W3105587016C178650346 @default.
- W3105587016 hasConceptScore W3105587016C19499675 @default.
- W3105587016 hasConceptScore W3105587016C202213908 @default.
- W3105587016 hasConceptScore W3105587016C202444582 @default.
- W3105587016 hasConceptScore W3105587016C2524010 @default.
- W3105587016 hasConceptScore W3105587016C2776214188 @default.
- W3105587016 hasConceptScore W3105587016C28826006 @default.
- W3105587016 hasConceptScore W3105587016C33923547 @default.
- W3105587016 hasConceptScore W3105587016C41008148 @default.
- W3105587016 hasConceptScore W3105587016C50644808 @default.
- W3105587016 hasConceptScore W3105587016C57830394 @default.
- W3105587016 hasConceptScore W3105587016C62520636 @default.
- W3105587016 hasConceptScore W3105587016C9652623 @default.
- W3105587016 hasLocation W31055870161 @default.
- W3105587016 hasOpenAccess W3105587016 @default.
- W3105587016 hasPrimaryLocation W31055870161 @default.
- W3105587016 hasRelatedWork W1788616518 @default.
- W3105587016 hasRelatedWork W2276001518 @default.
- W3105587016 hasRelatedWork W2372299347 @default.
- W3105587016 hasRelatedWork W2581090720 @default.
- W3105587016 hasRelatedWork W2611220308 @default.
- W3105587016 hasRelatedWork W2752086744 @default.
- W3105587016 hasRelatedWork W2769523825 @default.
- W3105587016 hasRelatedWork W2963303935 @default.
- W3105587016 hasRelatedWork W2964059111 @default.
- W3105587016 hasRelatedWork W2964250089 @default.
- W3105587016 hasRelatedWork W3019324987 @default.
- W3105587016 hasRelatedWork W3036446794 @default.
- W3105587016 hasRelatedWork W3040859283 @default.
- W3105587016 hasRelatedWork W3106195168 @default.
- W3105587016 hasRelatedWork W3123510000 @default.
- W3105587016 hasRelatedWork W3128023631 @default.
- W3105587016 hasRelatedWork W3155395043 @default.
- W3105587016 hasRelatedWork W3168194744 @default.
- W3105587016 hasRelatedWork W3170242741 @default.
- W3105587016 hasRelatedWork W599376124 @default.
- W3105587016 hasVolume "33" @default.
- W3105587016 isParatext "false" @default.
- W3105587016 isRetracted "false" @default.
- W3105587016 magId "3105587016" @default.
- W3105587016 workType "article" @default.