Matches in SemOpenAlex for { <https://semopenalex.org/work/W2912182380> ?p ?o ?g. }
- W2912182380 abstract "We propose a new variational family for Bayesian neural networks. We decompose the variational posterior into two components, where the radial component captures the strength of each neuron in terms of its magnitude; while the directional component captures the statistical dependencies among the weight parameters. The dependencies learned via the directional density provide better modeling performance compared to the widely-used Gaussian mean-field-type variational family. In addition, the strength of input and output neurons learned via the radial density provides a structured way to compress neural networks. Indeed, experiments show that our variational family improves predictive performance and yields compressed networks simultaneously." @default.
- W2912182380 created "2019-02-21" @default.
- W2912182380 creator A5015914441 @default.
- W2912182380 creator A5053558938 @default.
- W2912182380 creator A5081148743 @default.
- W2912182380 date "2019-02-07" @default.
- W2912182380 modified "2023-09-23" @default.
- W2912182380 title "Radial and Directional Posteriors for Bayesian Neural Networks." @default.
- W2912182380 cites W1520875569 @default.
- W2912182380 cites W1719489212 @default.
- W2912182380 cites W1826234144 @default.
- W2912182380 cites W1959608418 @default.
- W2912182380 cites W1983428013 @default.
- W2912182380 cites W1994841627 @default.
- W2912182380 cites W2042851446 @default.
- W2912182380 cites W2051125916 @default.
- W2912182380 cites W2081338783 @default.
- W2912182380 cites W2090406143 @default.
- W2912182380 cites W2108677974 @default.
- W2912182380 cites W2121593555 @default.
- W2912182380 cites W2122401230 @default.
- W2912182380 cites W2125389748 @default.
- W2912182380 cites W2139701068 @default.
- W2912182380 cites W2145001205 @default.
- W2912182380 cites W2158751090 @default.
- W2912182380 cites W2166851633 @default.
- W2912182380 cites W2257113116 @default.
- W2912182380 cites W2302053044 @default.
- W2912182380 cites W2402144811 @default.
- W2912182380 cites W2520760693 @default.
- W2912182380 cites W2547875792 @default.
- W2912182380 cites W2582745083 @default.
- W2912182380 cites W2602076750 @default.
- W2912182380 cites W2619890685 @default.
- W2912182380 cites W2701010355 @default.
- W2912182380 cites W2731982447 @default.
- W2912182380 cites W2786857698 @default.
- W2912182380 cites W2796335507 @default.
- W2912182380 cites W2899771611 @default.
- W2912182380 cites W2899948949 @default.
- W2912182380 cites W2951266961 @default.
- W2912182380 cites W2952165242 @default.
- W2912182380 cites W2952746978 @default.
- W2912182380 cites W2953263857 @default.
- W2912182380 cites W2962739609 @default.
- W2912182380 cites W2962965870 @default.
- W2912182380 cites W2963000224 @default.
- W2912182380 cites W2963117513 @default.
- W2912182380 cites W2963557175 @default.
- W2912182380 cites W2963711523 @default.
- W2912182380 cites W2964003575 @default.
- W2912182380 cites W2964059111 @default.
- W2912182380 cites W2964121744 @default.
- W2912182380 cites W2964260519 @default.
- W2912182380 cites W2964299589 @default.
- W2912182380 cites W566555209 @default.
- W2912182380 cites W992687842 @default.
- W2912182380 hasPublicationYear "2019" @default.
- W2912182380 type Work @default.
- W2912182380 sameAs 2912182380 @default.
- W2912182380 citedByCount "6" @default.
- W2912182380 countsByYear W29121823802019 @default.
- W2912182380 countsByYear W29121823802020 @default.
- W2912182380 countsByYear W29121823802021 @default.
- W2912182380 crossrefType "posted-content" @default.
- W2912182380 hasAuthorship W2912182380A5015914441 @default.
- W2912182380 hasAuthorship W2912182380A5053558938 @default.
- W2912182380 hasAuthorship W2912182380A5081148743 @default.
- W2912182380 hasConcept C107673813 @default.
- W2912182380 hasConcept C11413529 @default.
- W2912182380 hasConcept C121332964 @default.
- W2912182380 hasConcept C154945302 @default.
- W2912182380 hasConcept C163716315 @default.
- W2912182380 hasConcept C168167062 @default.
- W2912182380 hasConcept C33923547 @default.
- W2912182380 hasConcept C41008148 @default.
- W2912182380 hasConcept C50644808 @default.
- W2912182380 hasConcept C62520636 @default.
- W2912182380 hasConcept C97355855 @default.
- W2912182380 hasConceptScore W2912182380C107673813 @default.
- W2912182380 hasConceptScore W2912182380C11413529 @default.
- W2912182380 hasConceptScore W2912182380C121332964 @default.
- W2912182380 hasConceptScore W2912182380C154945302 @default.
- W2912182380 hasConceptScore W2912182380C163716315 @default.
- W2912182380 hasConceptScore W2912182380C168167062 @default.
- W2912182380 hasConceptScore W2912182380C33923547 @default.
- W2912182380 hasConceptScore W2912182380C41008148 @default.
- W2912182380 hasConceptScore W2912182380C50644808 @default.
- W2912182380 hasConceptScore W2912182380C62520636 @default.
- W2912182380 hasConceptScore W2912182380C97355855 @default.
- W2912182380 hasOpenAccess W2912182380 @default.
- W2912182380 hasRelatedWork W1498507519 @default.
- W2912182380 hasRelatedWork W1504805411 @default.
- W2912182380 hasRelatedWork W1554870656 @default.
- W2912182380 hasRelatedWork W1590360730 @default.
- W2912182380 hasRelatedWork W1783132139 @default.
- W2912182380 hasRelatedWork W1837230962 @default.
- W2912182380 hasRelatedWork W1961863157 @default.
- W2912182380 hasRelatedWork W1980290744 @default.
- W2912182380 hasRelatedWork W2095191728 @default.