Matches in SemOpenAlex for { <https://semopenalex.org/work/W2808333162> ?p ?o ?g. }
- W2808333162 abstract "Deep Gaussian Processes (DGPs) are hierarchical generalizations of Gaussian Processes that combine well calibrated uncertainty estimates with the high flexibility of multilayer models. One of the biggest challenges with these models is that exact inference is intractable. The current state-of-the-art inference method, Variational Inference (VI), employs a Gaussian approximation to the posterior distribution. This can be a potentially poor unimodal approximation of the generally multimodal posterior. In this work, we provide evidence for the non-Gaussian nature of the posterior and we apply the Stochastic Gradient Hamiltonian Monte Carlo method to generate samples. To efficiently optimize the hyperparameters, we introduce the Moving Window MCEM algorithm. This results in significantly better predictions at a lower computational cost than its VI counterpart. Thus our method establishes a new state-of-the-art for inference in DGPs." @default.
- W2808333162 created "2018-06-21" @default.
- W2808333162 creator A5029396252 @default.
- W2808333162 creator A5036960709 @default.
- W2808333162 creator A5091422392 @default.
- W2808333162 date "2018-06-14" @default.
- W2808333162 modified "2023-09-30" @default.
- W2808333162 title "Inference in Deep Gaussian Processes using Stochastic Gradient Hamiltonian Monte Carlo" @default.
- W2808333162 cites W1172736100 @default.
- W2808333162 cites W137285897 @default.
- W2808333162 cites W1534417711 @default.
- W2808333162 cites W1571870753 @default.
- W2808333162 cites W1577535158 @default.
- W2808333162 cites W1719489212 @default.
- W2808333162 cites W1934021597 @default.
- W2808333162 cites W195465510 @default.
- W2808333162 cites W2099768828 @default.
- W2808333162 cites W2108677974 @default.
- W2808333162 cites W2113571038 @default.
- W2808333162 cites W2134164499 @default.
- W2808333162 cites W2144193737 @default.
- W2808333162 cites W2150898568 @default.
- W2808333162 cites W2161767008 @default.
- W2808333162 cites W2167433878 @default.
- W2808333162 cites W2169410692 @default.
- W2808333162 cites W2257113116 @default.
- W2808333162 cites W2556372419 @default.
- W2808333162 cites W2559331598 @default.
- W2808333162 cites W2739517340 @default.
- W2808333162 cites W2774410544 @default.
- W2808333162 cites W2952372065 @default.
- W2808333162 cites W2953263857 @default.
- W2808333162 cites W2962875063 @default.
- W2808333162 cites W2963711523 @default.
- W2808333162 cites W2963717915 @default.
- W2808333162 cites W621546036 @default.
- W2808333162 cites W2072634211 @default.
- W2808333162 hasPublicationYear "2018" @default.
- W2808333162 type Work @default.
- W2808333162 sameAs 2808333162 @default.
- W2808333162 citedByCount "5" @default.
- W2808333162 countsByYear W28083331622019 @default.
- W2808333162 countsByYear W28083331622020 @default.
- W2808333162 countsByYear W28083331622021 @default.
- W2808333162 crossrefType "posted-content" @default.
- W2808333162 hasAuthorship W2808333162A5029396252 @default.
- W2808333162 hasAuthorship W2808333162A5036960709 @default.
- W2808333162 hasAuthorship W2808333162A5091422392 @default.
- W2808333162 hasConcept C105795698 @default.
- W2808333162 hasConcept C107673813 @default.
- W2808333162 hasConcept C111350023 @default.
- W2808333162 hasConcept C11413529 @default.
- W2808333162 hasConcept C121332964 @default.
- W2808333162 hasConcept C121864883 @default.
- W2808333162 hasConcept C126255220 @default.
- W2808333162 hasConcept C130787639 @default.
- W2808333162 hasConcept C13153151 @default.
- W2808333162 hasConcept C154945302 @default.
- W2808333162 hasConcept C163716315 @default.
- W2808333162 hasConcept C19499675 @default.
- W2808333162 hasConcept C2776214188 @default.
- W2808333162 hasConcept C2777472644 @default.
- W2808333162 hasConcept C28826006 @default.
- W2808333162 hasConcept C33923547 @default.
- W2808333162 hasConcept C41008148 @default.
- W2808333162 hasConcept C57830394 @default.
- W2808333162 hasConcept C61326573 @default.
- W2808333162 hasConcept C62520636 @default.
- W2808333162 hasConcept C8642999 @default.
- W2808333162 hasConceptScore W2808333162C105795698 @default.
- W2808333162 hasConceptScore W2808333162C107673813 @default.
- W2808333162 hasConceptScore W2808333162C111350023 @default.
- W2808333162 hasConceptScore W2808333162C11413529 @default.
- W2808333162 hasConceptScore W2808333162C121332964 @default.
- W2808333162 hasConceptScore W2808333162C121864883 @default.
- W2808333162 hasConceptScore W2808333162C126255220 @default.
- W2808333162 hasConceptScore W2808333162C130787639 @default.
- W2808333162 hasConceptScore W2808333162C13153151 @default.
- W2808333162 hasConceptScore W2808333162C154945302 @default.
- W2808333162 hasConceptScore W2808333162C163716315 @default.
- W2808333162 hasConceptScore W2808333162C19499675 @default.
- W2808333162 hasConceptScore W2808333162C2776214188 @default.
- W2808333162 hasConceptScore W2808333162C2777472644 @default.
- W2808333162 hasConceptScore W2808333162C28826006 @default.
- W2808333162 hasConceptScore W2808333162C33923547 @default.
- W2808333162 hasConceptScore W2808333162C41008148 @default.
- W2808333162 hasConceptScore W2808333162C57830394 @default.
- W2808333162 hasConceptScore W2808333162C61326573 @default.
- W2808333162 hasConceptScore W2808333162C62520636 @default.
- W2808333162 hasConceptScore W2808333162C8642999 @default.
- W2808333162 hasLocation W28083331621 @default.
- W2808333162 hasOpenAccess W2808333162 @default.
- W2808333162 hasPrimaryLocation W28083331621 @default.
- W2808333162 hasRelatedWork W176758818 @default.
- W2808333162 hasRelatedWork W177540543 @default.
- W2808333162 hasRelatedWork W1829113803 @default.
- W2808333162 hasRelatedWork W1927271208 @default.
- W2808333162 hasRelatedWork W2117804315 @default.
- W2808333162 hasRelatedWork W2156417670 @default.
- W2808333162 hasRelatedWork W2243234527 @default.