FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3024650197> ?p ?o ?g. }

• W3024650197 abstract "We consider the optimal approximate posterior over the top-layer weights in a Bayesian neural network for regression, and show that it exhibits strong dependencies on the lower-layer weights. We adapt this result to develop a correlated approximate posterior over the weights at all layers in a Bayesian neural network. We extend this approach to deep Gaussian processes, unifying inference in the two model classes. Our approximate posterior uses learned global inducing points, which are defined only at the input layer and propagated through the network to obtain inducing inputs at subsequent layers. By contrast, standard, local, inducing point methods from the deep Gaussian process literature optimise a separate set of inducing inputs at every layer, and thus do not model correlations across layers. Our method gives state-of-the-art performance for a variational Bayesian method, without data augmentation or tempering, on CIFAR-10 of 86.7%, which is comparable to SGMCMC without tempering but with data augmentation (88% in Wenzel et al. 2020)." @default.
• W3024650197 created "2020-05-21" @default.
• W3024650197 creator A5057921753 @default.
• W3024650197 creator A5062539506 @default.
• W3024650197 date "2020-05-17" @default.
• W3024650197 modified "2023-09-27" @default.
• W3024650197 title "Global inducing point variational posteriors for Bayesian neural networks and deep Gaussian processes" @default.
• W3024650197 cites W137285897 @default.
• W3024650197 cites W1516111018 @default.
• W3024650197 cites W1522301498 @default.
• W3024650197 cites W1677182931 @default.
• W3024650197 cites W1719489212 @default.
• W3024650197 cites W1795258949 @default.
• W3024650197 cites W1826234144 @default.
• W3024650197 cites W1836465849 @default.
• W3024650197 cites W1909320841 @default.
• W3024650197 cites W2047229728 @default.
• W3024650197 cites W2099768828 @default.
• W3024650197 cites W2108677974 @default.
• W3024650197 cites W215707797 @default.
• W3024650197 cites W2159080219 @default.
• W3024650197 cites W2194775991 @default.
• W3024650197 cites W2254249950 @default.
• W3024650197 cites W2302053044 @default.
• W3024650197 cites W2335728318 @default.
• W3024650197 cites W2478027467 @default.
• W3024650197 cites W2592505114 @default.
• W3024650197 cites W2619650475 @default.
• W3024650197 cites W2753141617 @default.
• W3024650197 cites W2766678531 @default.
• W3024650197 cites W2767449908 @default.
• W3024650197 cites W2774412855 @default.
• W3024650197 cites W2784590060 @default.
• W3024650197 cites W2785626633 @default.
• W3024650197 cites W2786857698 @default.
• W3024650197 cites W2902363477 @default.
• W3024650197 cites W2910171846 @default.
• W3024650197 cites W2911315040 @default.
• W3024650197 cites W2951004968 @default.
• W3024650197 cites W2951266961 @default.
• W3024650197 cites W2952816888 @default.
• W3024650197 cites W2955023002 @default.
• W3024650197 cites W2962875063 @default.
• W3024650197 cites W2963505303 @default.
• W3024650197 cites W2963711523 @default.
• W3024650197 cites W2964052395 @default.
• W3024650197 cites W2964059111 @default.
• W3024650197 cites W2964212410 @default.
• W3024650197 cites W2970861023 @default.
• W3024650197 cites W2970920715 @default.
• W3024650197 cites W2971941774 @default.
• W3024650197 cites W2980558348 @default.
• W3024650197 cites W2987473824 @default.
• W3024650197 cites W2999506416 @default.
• W3024650197 cites W3004944559 @default.
• W3024650197 cites W3005141979 @default.
• W3024650197 cites W3005727684 @default.
• W3024650197 cites W3006511999 @default.
• W3024650197 cites W3026026123 @default.
• W3024650197 cites W3026126175 @default.
• W3024650197 cites W3034669169 @default.
• W3024650197 cites W3037873282 @default.
• W3024650197 cites W3040859283 @default.
• W3024650197 cites W3048657615 @default.
• W3024650197 cites W3081084362 @default.
• W3024650197 cites W3089559016 @default.
• W3024650197 cites W3118608800 @default.
• W3024650197 cites W3130833330 @default.
• W3024650197 cites W3131195611 @default.
• W3024650197 cites W3166861954 @default.
• W3024650197 hasPublicationYear "2020" @default.
• W3024650197 type Work @default.
• W3024650197 sameAs 3024650197 @default.
• W3024650197 citedByCount "10" @default.
• W3024650197 countsByYear W30246501972020 @default.
• W3024650197 countsByYear W30246501972021 @default.
• W3024650197 crossrefType "posted-content" @default.
• W3024650197 hasAuthorship W3024650197A5057921753 @default.
• W3024650197 hasAuthorship W3024650197A5062539506 @default.
• W3024650197 hasConcept C107673813 @default.
• W3024650197 hasConcept C11413529 @default.
• W3024650197 hasConcept C121332964 @default.
• W3024650197 hasConcept C154945302 @default.
• W3024650197 hasConcept C160234255 @default.
• W3024650197 hasConcept C163716315 @default.
• W3024650197 hasConcept C177264268 @default.
• W3024650197 hasConcept C178790620 @default.
• W3024650197 hasConcept C185592680 @default.
• W3024650197 hasConcept C199360897 @default.
• W3024650197 hasConcept C2524010 @default.
• W3024650197 hasConcept C2776214188 @default.
• W3024650197 hasConcept C2779227376 @default.
• W3024650197 hasConcept C28719098 @default.
• W3024650197 hasConcept C33923547 @default.
• W3024650197 hasConcept C41008148 @default.
• W3024650197 hasConcept C50644808 @default.
• W3024650197 hasConcept C57830394 @default.
• W3024650197 hasConcept C61326573 @default.
• W3024650197 hasConcept C62520636 @default.
• W3024650197 hasConceptScore W3024650197C107673813 @default.

• next