Matches in SemOpenAlex for { <https://semopenalex.org/work/W2978503189> ?p ?o ?g. }
Showing items 1 to 89 of
89
with 100 items per page.
- W2978503189 abstract "The Laplace approximation has been one of the workhorses of Bayesian inference. It often delivers good approximations in practice despite the fact that it does not strictly take into account where the volume of posterior density lies. Variational approaches avoid this issue by explicitly minimising the Kullback-Leibler divergence DKL between a postulated posterior and the true (unnormalised) logarithmic posterior. However, they rely on a closed form DKL in order to update the variational parameters. To address this, stochastic versions of variational inference have been devised that approximate the intractable DKL with a Monte Carlo average. This approximation allows calculating gradients with respect to the variational parameters. However, variational methods often postulate a factorised Gaussian approximating posterior. In doing so, they sacrifice a-posteriori correlations. In this work, we propose a method that combines the Laplace approximation with the variational approach. The advantages are that we maintain: applicability on non-conjugate models, posterior correlations and a reduced number of free variational parameters. Numerical experiments demonstrate improvement over the Laplace approximation and variational inference with factorised Gaussian posteriors." @default.
- W2978503189 created "2019-10-10" @default.
- W2978503189 creator A5060507310 @default.
- W2978503189 date "2019-07-01" @default.
- W2978503189 modified "2023-10-16" @default.
- W2978503189 title "Mixed Variational Inference" @default.
- W2978503189 cites W1515605016 @default.
- W2978503189 cites W2000721204 @default.
- W2978503189 cites W2109717317 @default.
- W2978503189 cites W2111051539 @default.
- W2978503189 cites W2157525865 @default.
- W2978503189 cites W2166361453 @default.
- W2978503189 cites W2167234030 @default.
- W2978503189 cites W2169228078 @default.
- W2978503189 cites W2223412984 @default.
- W2978503189 cites W2796079430 @default.
- W2978503189 cites W3100197521 @default.
- W2978503189 cites W3104819538 @default.
- W2978503189 cites W4248681815 @default.
- W2978503189 cites W4292199333 @default.
- W2978503189 cites W4298872162 @default.
- W2978503189 doi "https://doi.org/10.1109/ijcnn.2019.8852348" @default.
- W2978503189 hasPublicationYear "2019" @default.
- W2978503189 type Work @default.
- W2978503189 sameAs 2978503189 @default.
- W2978503189 citedByCount "0" @default.
- W2978503189 crossrefType "proceedings-article" @default.
- W2978503189 hasAuthorship W2978503189A5060507310 @default.
- W2978503189 hasBestOaLocation W29785031892 @default.
- W2978503189 hasConcept C105795698 @default.
- W2978503189 hasConcept C107673813 @default.
- W2978503189 hasConcept C121332964 @default.
- W2978503189 hasConcept C121864883 @default.
- W2978503189 hasConcept C126255220 @default.
- W2978503189 hasConcept C134306372 @default.
- W2978503189 hasConcept C138885662 @default.
- W2978503189 hasConcept C154945302 @default.
- W2978503189 hasConcept C160234255 @default.
- W2978503189 hasConcept C163716315 @default.
- W2978503189 hasConcept C19499675 @default.
- W2978503189 hasConcept C207390915 @default.
- W2978503189 hasConcept C22243797 @default.
- W2978503189 hasConcept C2776214188 @default.
- W2978503189 hasConcept C28826006 @default.
- W2978503189 hasConcept C33923547 @default.
- W2978503189 hasConcept C41008148 @default.
- W2978503189 hasConcept C41895202 @default.
- W2978503189 hasConcept C57830394 @default.
- W2978503189 hasConcept C62520636 @default.
- W2978503189 hasConcept C97937538 @default.
- W2978503189 hasConceptScore W2978503189C105795698 @default.
- W2978503189 hasConceptScore W2978503189C107673813 @default.
- W2978503189 hasConceptScore W2978503189C121332964 @default.
- W2978503189 hasConceptScore W2978503189C121864883 @default.
- W2978503189 hasConceptScore W2978503189C126255220 @default.
- W2978503189 hasConceptScore W2978503189C134306372 @default.
- W2978503189 hasConceptScore W2978503189C138885662 @default.
- W2978503189 hasConceptScore W2978503189C154945302 @default.
- W2978503189 hasConceptScore W2978503189C160234255 @default.
- W2978503189 hasConceptScore W2978503189C163716315 @default.
- W2978503189 hasConceptScore W2978503189C19499675 @default.
- W2978503189 hasConceptScore W2978503189C207390915 @default.
- W2978503189 hasConceptScore W2978503189C22243797 @default.
- W2978503189 hasConceptScore W2978503189C2776214188 @default.
- W2978503189 hasConceptScore W2978503189C28826006 @default.
- W2978503189 hasConceptScore W2978503189C33923547 @default.
- W2978503189 hasConceptScore W2978503189C41008148 @default.
- W2978503189 hasConceptScore W2978503189C41895202 @default.
- W2978503189 hasConceptScore W2978503189C57830394 @default.
- W2978503189 hasConceptScore W2978503189C62520636 @default.
- W2978503189 hasConceptScore W2978503189C97937538 @default.
- W2978503189 hasLocation W29785031891 @default.
- W2978503189 hasLocation W29785031892 @default.
- W2978503189 hasOpenAccess W2978503189 @default.
- W2978503189 hasPrimaryLocation W29785031891 @default.
- W2978503189 hasRelatedWork W2138213559 @default.
- W2978503189 hasRelatedWork W2188349433 @default.
- W2978503189 hasRelatedWork W2235729179 @default.
- W2978503189 hasRelatedWork W2396956736 @default.
- W2978503189 hasRelatedWork W2531556462 @default.
- W2978503189 hasRelatedWork W2574967310 @default.
- W2978503189 hasRelatedWork W2581554909 @default.
- W2978503189 hasRelatedWork W2795572003 @default.
- W2978503189 hasRelatedWork W2949221364 @default.
- W2978503189 hasRelatedWork W3168765527 @default.
- W2978503189 isParatext "false" @default.
- W2978503189 isRetracted "false" @default.
- W2978503189 magId "2978503189" @default.
- W2978503189 workType "article" @default.