Matches in SemOpenAlex for { <https://semopenalex.org/work/W4281764037> ?p ?o ?g. }
Showing items 1 to 71 of
71
with 100 items per page.
- W4281764037 abstract "Sampling from complex target distributions is a challenging task fundamental to Bayesian inference. Parallel tempering (PT) addresses this problem by constructing a Markov chain on the expanded state space of a sequence of distributions interpolating between the posterior distribution and a fixed reference distribution, which is typically chosen to be the prior. However, in the typical case where the prior and posterior are nearly mutually singular, PT methods are computationally prohibitive. In this work we address this challenge by constructing a generalized annealing path connecting the posterior to an adaptively tuned variational reference. The reference distribution is tuned to minimize the forward (inclusive) KL divergence to the posterior distribution using a simple, gradient-free moment-matching procedure. We show that our adaptive procedure converges to the forward KL minimizer, and that the forward KL divergence serves as a good proxy to a previously developed measure of PT performance. We also show that in the large-data limit in typical Bayesian models, the proposed method improves in performance, while traditional PT deteriorates arbitrarily. Finally, we introduce PT with two references -- one fixed, one variational -- with a novel split annealing path that ensures stable variational reference adaptation. The paper concludes with experiments that demonstrate the large empirical gains achieved by our method in a wide range of realistic Bayesian inference scenarios." @default.
- W4281764037 created "2022-06-13" @default.
- W4281764037 creator A5003160514 @default.
- W4281764037 creator A5003426513 @default.
- W4281764037 creator A5004323430 @default.
- W4281764037 creator A5023079216 @default.
- W4281764037 date "2022-05-31" @default.
- W4281764037 modified "2023-09-29" @default.
- W4281764037 title "Parallel Tempering With a Variational Reference" @default.
- W4281764037 doi "https://doi.org/10.48550/arxiv.2206.00080" @default.
- W4281764037 hasPublicationYear "2022" @default.
- W4281764037 type Work @default.
- W4281764037 citedByCount "0" @default.
- W4281764037 crossrefType "posted-content" @default.
- W4281764037 hasAuthorship W4281764037A5003160514 @default.
- W4281764037 hasAuthorship W4281764037A5003426513 @default.
- W4281764037 hasAuthorship W4281764037A5004323430 @default.
- W4281764037 hasAuthorship W4281764037A5023079216 @default.
- W4281764037 hasBestOaLocation W42817640371 @default.
- W4281764037 hasConcept C107673813 @default.
- W4281764037 hasConcept C111350023 @default.
- W4281764037 hasConcept C11413529 @default.
- W4281764037 hasConcept C119857082 @default.
- W4281764037 hasConcept C126255220 @default.
- W4281764037 hasConcept C126980161 @default.
- W4281764037 hasConcept C138885662 @default.
- W4281764037 hasConcept C154945302 @default.
- W4281764037 hasConcept C160234255 @default.
- W4281764037 hasConcept C187653413 @default.
- W4281764037 hasConcept C207390915 @default.
- W4281764037 hasConcept C28826006 @default.
- W4281764037 hasConcept C33923547 @default.
- W4281764037 hasConcept C37669827 @default.
- W4281764037 hasConcept C41008148 @default.
- W4281764037 hasConcept C41895202 @default.
- W4281764037 hasConcept C57830394 @default.
- W4281764037 hasConcept C98763669 @default.
- W4281764037 hasConceptScore W4281764037C107673813 @default.
- W4281764037 hasConceptScore W4281764037C111350023 @default.
- W4281764037 hasConceptScore W4281764037C11413529 @default.
- W4281764037 hasConceptScore W4281764037C119857082 @default.
- W4281764037 hasConceptScore W4281764037C126255220 @default.
- W4281764037 hasConceptScore W4281764037C126980161 @default.
- W4281764037 hasConceptScore W4281764037C138885662 @default.
- W4281764037 hasConceptScore W4281764037C154945302 @default.
- W4281764037 hasConceptScore W4281764037C160234255 @default.
- W4281764037 hasConceptScore W4281764037C187653413 @default.
- W4281764037 hasConceptScore W4281764037C207390915 @default.
- W4281764037 hasConceptScore W4281764037C28826006 @default.
- W4281764037 hasConceptScore W4281764037C33923547 @default.
- W4281764037 hasConceptScore W4281764037C37669827 @default.
- W4281764037 hasConceptScore W4281764037C41008148 @default.
- W4281764037 hasConceptScore W4281764037C41895202 @default.
- W4281764037 hasConceptScore W4281764037C57830394 @default.
- W4281764037 hasConceptScore W4281764037C98763669 @default.
- W4281764037 hasLocation W42817640371 @default.
- W4281764037 hasOpenAccess W4281764037 @default.
- W4281764037 hasPrimaryLocation W42817640371 @default.
- W4281764037 hasRelatedWork W2013478520 @default.
- W4281764037 hasRelatedWork W2027276540 @default.
- W4281764037 hasRelatedWork W2033357194 @default.
- W4281764037 hasRelatedWork W2060098566 @default.
- W4281764037 hasRelatedWork W2114058417 @default.
- W4281764037 hasRelatedWork W2366775747 @default.
- W4281764037 hasRelatedWork W2790979771 @default.
- W4281764037 hasRelatedWork W3086697448 @default.
- W4281764037 hasRelatedWork W3087071515 @default.
- W4281764037 hasRelatedWork W3103384301 @default.
- W4281764037 isParatext "false" @default.
- W4281764037 isRetracted "false" @default.
- W4281764037 workType "article" @default.