Matches in SemOpenAlex for { <https://semopenalex.org/work/W3007431310> ?p ?o ?g. }
Showing items 1 to 78 of
78
with 100 items per page.
- W3007431310 abstract "A common divide-and-conquer approach for Bayesian computation with big data is to partition the data, perform local inference for each piece separately, and combine the results to obtain a global posterior approximation. While being conceptually and computationally appealing, this method involves the problematic need to also split the prior for the local inferences; these weakened priors may not provide enough regularization for each separate computation, thus eliminating one of the key advantages of Bayesian methods. To resolve this dilemma while still retaining the generalizability of the underlying local inference method, we apply the idea of expectation propagation (EP) as a framework for distributed Bayesian inference. The central idea is to iteratively update approximations to the local likelihoods given the state of the other approximations and the prior. The present paper has two roles: we review the steps that are needed to keep EP algorithms numerically stable, and we suggest a general approach, inspired by EP, for approaching data partitioning problems in a way that achieves the computational benefits of parallelism while allowing each local update to make use of relevant information from the other sites. In addition, we demonstrate how the method can be applied in a hierarchical context to make use of partitioning of both data and parameters. The paper describes a general algorithmic framework, rather than a specific algorithm, and presents an example implementation for it." @default.
- W3007431310 created "2020-03-06" @default.
- W3007431310 creator A5000113707 @default.
- W3007431310 creator A5006293834 @default.
- W3007431310 creator A5013861260 @default.
- W3007431310 creator A5030697996 @default.
- W3007431310 creator A5047269552 @default.
- W3007431310 creator A5048712208 @default.
- W3007431310 creator A5074093929 @default.
- W3007431310 creator A5074835142 @default.
- W3007431310 creator A5080652581 @default.
- W3007431310 creator A5081819708 @default.
- W3007431310 date "2020-12-01" @default.
- W3007431310 modified "2023-09-25" @default.
- W3007431310 title "Expectation Propagation as a Way of Life: A Framework for Bayesian Inference on Partitioned Data" @default.
- W3007431310 hasPublicationYear "2020" @default.
- W3007431310 type Work @default.
- W3007431310 sameAs 3007431310 @default.
- W3007431310 citedByCount "12" @default.
- W3007431310 countsByYear W30074313102015 @default.
- W3007431310 countsByYear W30074313102017 @default.
- W3007431310 countsByYear W30074313102019 @default.
- W3007431310 countsByYear W30074313102020 @default.
- W3007431310 countsByYear W30074313102021 @default.
- W3007431310 crossrefType "journal-article" @default.
- W3007431310 hasAuthorship W3007431310A5000113707 @default.
- W3007431310 hasAuthorship W3007431310A5006293834 @default.
- W3007431310 hasAuthorship W3007431310A5013861260 @default.
- W3007431310 hasAuthorship W3007431310A5030697996 @default.
- W3007431310 hasAuthorship W3007431310A5047269552 @default.
- W3007431310 hasAuthorship W3007431310A5048712208 @default.
- W3007431310 hasAuthorship W3007431310A5074093929 @default.
- W3007431310 hasAuthorship W3007431310A5074835142 @default.
- W3007431310 hasAuthorship W3007431310A5080652581 @default.
- W3007431310 hasAuthorship W3007431310A5081819708 @default.
- W3007431310 hasBestOaLocation W30074313102 @default.
- W3007431310 hasConcept C107673813 @default.
- W3007431310 hasConcept C11413529 @default.
- W3007431310 hasConcept C119857082 @default.
- W3007431310 hasConcept C154945302 @default.
- W3007431310 hasConcept C160234255 @default.
- W3007431310 hasConcept C177769412 @default.
- W3007431310 hasConcept C2776135515 @default.
- W3007431310 hasConcept C2776214188 @default.
- W3007431310 hasConcept C2779377595 @default.
- W3007431310 hasConcept C41008148 @default.
- W3007431310 hasConcept C45374587 @default.
- W3007431310 hasConcept C80444323 @default.
- W3007431310 hasConceptScore W3007431310C107673813 @default.
- W3007431310 hasConceptScore W3007431310C11413529 @default.
- W3007431310 hasConceptScore W3007431310C119857082 @default.
- W3007431310 hasConceptScore W3007431310C154945302 @default.
- W3007431310 hasConceptScore W3007431310C160234255 @default.
- W3007431310 hasConceptScore W3007431310C177769412 @default.
- W3007431310 hasConceptScore W3007431310C2776135515 @default.
- W3007431310 hasConceptScore W3007431310C2776214188 @default.
- W3007431310 hasConceptScore W3007431310C2779377595 @default.
- W3007431310 hasConceptScore W3007431310C41008148 @default.
- W3007431310 hasConceptScore W3007431310C45374587 @default.
- W3007431310 hasConceptScore W3007431310C80444323 @default.
- W3007431310 hasLocation W30074313101 @default.
- W3007431310 hasLocation W30074313102 @default.
- W3007431310 hasOpenAccess W3007431310 @default.
- W3007431310 hasPrimaryLocation W30074313101 @default.
- W3007431310 hasRelatedWork W1498794143 @default.
- W3007431310 hasRelatedWork W2901441036 @default.
- W3007431310 hasRelatedWork W2963058055 @default.
- W3007431310 hasRelatedWork W3124497799 @default.
- W3007431310 hasRelatedWork W3143447564 @default.
- W3007431310 hasRelatedWork W3165541948 @default.
- W3007431310 hasRelatedWork W4280548097 @default.
- W3007431310 hasRelatedWork W4289282342 @default.
- W3007431310 hasRelatedWork W4297744796 @default.
- W3007431310 hasRelatedWork W4307312013 @default.
- W3007431310 isParatext "false" @default.
- W3007431310 isRetracted "false" @default.
- W3007431310 magId "3007431310" @default.
- W3007431310 workType "article" @default.