Matches in SemOpenAlex for { <https://semopenalex.org/work/W3100488384> ?p ?o ?g. }
Showing items 1 to 80 of
80
with 100 items per page.
- W3100488384 endingPage "17152" @default.
- W3100488384 startingPage "17141" @default.
- W3100488384 abstract "Probabilistic neural networks are typically modeled with independent weight priors, which do not capture weight correlations in the prior and do not provide a parsimonious interface to express properties in function space. A desirable class of priors would represent weights compactly, capture correlations between weights, facilitate calibrated reasoning about uncertainty, and allow inclusion of prior knowledge about the function space such as periodicity or dependence on contexts such as inputs. To this end, this paper introduces two innovations: (i) a Gaussian process-based hierarchical model for network weights based on unit embeddings that can flexibly encode correlated weight structures, and (ii) input-dependent versions of these weight priors that can provide convenient ways to regularize the function space through the use of kernels defined on contextual inputs. We show these models provide desirable test-time uncertainty estimates on out-of-distribution data, demonstrate cases of modeling inductive biases for neural networks with kernels which help both interpolation and extrapolation from training data, and demonstrate competitive predictive performance on an active learning benchmark." @default.
- W3100488384 created "2020-11-23" @default.
- W3100488384 creator A5000188286 @default.
- W3100488384 creator A5064498575 @default.
- W3100488384 date "2020-02-10" @default.
- W3100488384 modified "2023-09-24" @default.
- W3100488384 title "Hierarchical Gaussian Process Priors for Bayesian Neural Network Weights" @default.
- W3100488384 hasPublicationYear "2020" @default.
- W3100488384 type Work @default.
- W3100488384 sameAs 3100488384 @default.
- W3100488384 citedByCount "8" @default.
- W3100488384 countsByYear W31004883842020 @default.
- W3100488384 countsByYear W31004883842021 @default.
- W3100488384 crossrefType "proceedings-article" @default.
- W3100488384 hasAuthorship W3100488384A5000188286 @default.
- W3100488384 hasAuthorship W3100488384A5064498575 @default.
- W3100488384 hasConcept C105795698 @default.
- W3100488384 hasConcept C107673813 @default.
- W3100488384 hasConcept C11413529 @default.
- W3100488384 hasConcept C119857082 @default.
- W3100488384 hasConcept C121332964 @default.
- W3100488384 hasConcept C132459708 @default.
- W3100488384 hasConcept C13280743 @default.
- W3100488384 hasConcept C154945302 @default.
- W3100488384 hasConcept C163716315 @default.
- W3100488384 hasConcept C177769412 @default.
- W3100488384 hasConcept C185798385 @default.
- W3100488384 hasConcept C205649164 @default.
- W3100488384 hasConcept C33923547 @default.
- W3100488384 hasConcept C41008148 @default.
- W3100488384 hasConcept C50644808 @default.
- W3100488384 hasConcept C61326573 @default.
- W3100488384 hasConcept C62520636 @default.
- W3100488384 hasConceptScore W3100488384C105795698 @default.
- W3100488384 hasConceptScore W3100488384C107673813 @default.
- W3100488384 hasConceptScore W3100488384C11413529 @default.
- W3100488384 hasConceptScore W3100488384C119857082 @default.
- W3100488384 hasConceptScore W3100488384C121332964 @default.
- W3100488384 hasConceptScore W3100488384C132459708 @default.
- W3100488384 hasConceptScore W3100488384C13280743 @default.
- W3100488384 hasConceptScore W3100488384C154945302 @default.
- W3100488384 hasConceptScore W3100488384C163716315 @default.
- W3100488384 hasConceptScore W3100488384C177769412 @default.
- W3100488384 hasConceptScore W3100488384C185798385 @default.
- W3100488384 hasConceptScore W3100488384C205649164 @default.
- W3100488384 hasConceptScore W3100488384C33923547 @default.
- W3100488384 hasConceptScore W3100488384C41008148 @default.
- W3100488384 hasConceptScore W3100488384C50644808 @default.
- W3100488384 hasConceptScore W3100488384C61326573 @default.
- W3100488384 hasConceptScore W3100488384C62520636 @default.
- W3100488384 hasLocation W31004883841 @default.
- W3100488384 hasOpenAccess W3100488384 @default.
- W3100488384 hasPrimaryLocation W31004883841 @default.
- W3100488384 hasRelatedWork W2126986674 @default.
- W3100488384 hasRelatedWork W2194775991 @default.
- W3100488384 hasRelatedWork W2592505114 @default.
- W3100488384 hasRelatedWork W2604977777 @default.
- W3100488384 hasRelatedWork W2755098895 @default.
- W3100488384 hasRelatedWork W2782487094 @default.
- W3100488384 hasRelatedWork W2899875051 @default.
- W3100488384 hasRelatedWork W2963238274 @default.
- W3100488384 hasRelatedWork W2970859221 @default.
- W3100488384 hasRelatedWork W2979661719 @default.
- W3100488384 hasRelatedWork W2981967650 @default.
- W3100488384 hasRelatedWork W3005727684 @default.
- W3100488384 hasRelatedWork W3034669169 @default.
- W3100488384 hasRelatedWork W3045583460 @default.
- W3100488384 hasRelatedWork W3093463114 @default.
- W3100488384 hasRelatedWork W3093515944 @default.
- W3100488384 hasRelatedWork W3130833330 @default.
- W3100488384 hasRelatedWork W3179857804 @default.
- W3100488384 hasRelatedWork W3200177339 @default.
- W3100488384 hasRelatedWork W3202239698 @default.
- W3100488384 hasVolume "33" @default.
- W3100488384 isParatext "false" @default.
- W3100488384 isRetracted "false" @default.
- W3100488384 magId "3100488384" @default.
- W3100488384 workType "article" @default.