Matches in SemOpenAlex for { <https://semopenalex.org/work/W3048497507> ?p ?o ?g. }
- W3048497507 abstract "The marginal likelihood or evidence in Bayesian statistics contains an intrinsic penalty for larger model sizes and is a fundamental quantity in Bayesian model comparison. Over the past two decades, there has been steadily increasing activity to understand the nature of this penalty in singular statistical models, building on pioneering work by Sumio Watanabe. Unlike regular models where the Bayesian information criterion (BIC) encapsulates a first-order expansion of the logarithm of the marginal likelihood, parameter counting gets trickier in singular models where a quantity called the real log canonical threshold (RLCT) summarizes the effective model dimensionality. In this article, we offer a probabilistic treatment to recover non-asymptotic versions of established evidence bounds as well as prove a new result based on the Gibbs variational inequality. In particular, we show that mean-field variational inference correctly recovers the RLCT for any singular model in its canonical or normal form. We additionally exhibit sharpness of our bound by analyzing the dynamics of a general purpose coordinate ascent algorithm (CAVI) popularly employed in variational inference." @default.
- W3048497507 created "2020-08-18" @default.
- W3048497507 creator A5033365412 @default.
- W3048497507 creator A5034514833 @default.
- W3048497507 creator A5075421264 @default.
- W3048497507 date "2020-08-11" @default.
- W3048497507 modified "2023-09-27" @default.
- W3048497507 title "Evidence bounds in singular models: probabilistic and variational perspectives" @default.
- W3048497507 cites W1501542638 @default.
- W3048497507 cites W1530226021 @default.
- W3048497507 cites W1546772535 @default.
- W3048497507 cites W1559340406 @default.
- W3048497507 cites W1594725708 @default.
- W3048497507 cites W1663973292 @default.
- W3048497507 cites W1968936725 @default.
- W3048497507 cites W2012644626 @default.
- W3048497507 cites W2015114311 @default.
- W3048497507 cites W2017696952 @default.
- W3048497507 cites W2023792765 @default.
- W3048497507 cites W2030085689 @default.
- W3048497507 cites W2082206048 @default.
- W3048497507 cites W2094438648 @default.
- W3048497507 cites W2103606110 @default.
- W3048497507 cites W2105304730 @default.
- W3048497507 cites W2106095534 @default.
- W3048497507 cites W2106568142 @default.
- W3048497507 cites W2107455764 @default.
- W3048497507 cites W2110009816 @default.
- W3048497507 cites W2117076645 @default.
- W3048497507 cites W2120217353 @default.
- W3048497507 cites W2120340025 @default.
- W3048497507 cites W2120609536 @default.
- W3048497507 cites W2125015779 @default.
- W3048497507 cites W2141200443 @default.
- W3048497507 cites W2168175751 @default.
- W3048497507 cites W2292751606 @default.
- W3048497507 cites W2785731189 @default.
- W3048497507 cites W2787248994 @default.
- W3048497507 cites W2801179766 @default.
- W3048497507 cites W2890494455 @default.
- W3048497507 cites W2948744601 @default.
- W3048497507 cites W2962774697 @default.
- W3048497507 cites W2962804724 @default.
- W3048497507 cites W2963742450 @default.
- W3048497507 cites W2964096787 @default.
- W3048497507 cites W2964306937 @default.
- W3048497507 cites W2990874482 @default.
- W3048497507 cites W3033456699 @default.
- W3048497507 cites W3123857276 @default.
- W3048497507 cites W3140968660 @default.
- W3048497507 cites W2585198796 @default.
- W3048497507 cites W3014551858 @default.
- W3048497507 hasPublicationYear "2020" @default.
- W3048497507 type Work @default.
- W3048497507 sameAs 3048497507 @default.
- W3048497507 citedByCount "0" @default.
- W3048497507 crossrefType "posted-content" @default.
- W3048497507 hasAuthorship W3048497507A5033365412 @default.
- W3048497507 hasAuthorship W3048497507A5034514833 @default.
- W3048497507 hasAuthorship W3048497507A5075421264 @default.
- W3048497507 hasConcept C105795698 @default.
- W3048497507 hasConcept C107673813 @default.
- W3048497507 hasConcept C111030470 @default.
- W3048497507 hasConcept C114289077 @default.
- W3048497507 hasConcept C134306372 @default.
- W3048497507 hasConcept C154945302 @default.
- W3048497507 hasConcept C160234255 @default.
- W3048497507 hasConcept C2776214188 @default.
- W3048497507 hasConcept C28826006 @default.
- W3048497507 hasConcept C33923547 @default.
- W3048497507 hasConcept C39927690 @default.
- W3048497507 hasConcept C41008148 @default.
- W3048497507 hasConcept C49937458 @default.
- W3048497507 hasConcept C95923904 @default.
- W3048497507 hasConceptScore W3048497507C105795698 @default.
- W3048497507 hasConceptScore W3048497507C107673813 @default.
- W3048497507 hasConceptScore W3048497507C111030470 @default.
- W3048497507 hasConceptScore W3048497507C114289077 @default.
- W3048497507 hasConceptScore W3048497507C134306372 @default.
- W3048497507 hasConceptScore W3048497507C154945302 @default.
- W3048497507 hasConceptScore W3048497507C160234255 @default.
- W3048497507 hasConceptScore W3048497507C2776214188 @default.
- W3048497507 hasConceptScore W3048497507C28826006 @default.
- W3048497507 hasConceptScore W3048497507C33923547 @default.
- W3048497507 hasConceptScore W3048497507C39927690 @default.
- W3048497507 hasConceptScore W3048497507C41008148 @default.
- W3048497507 hasConceptScore W3048497507C49937458 @default.
- W3048497507 hasConceptScore W3048497507C95923904 @default.
- W3048497507 hasLocation W30484975071 @default.
- W3048497507 hasOpenAccess W3048497507 @default.
- W3048497507 hasPrimaryLocation W30484975071 @default.
- W3048497507 hasRelatedWork W1545094706 @default.
- W3048497507 hasRelatedWork W1601059808 @default.
- W3048497507 hasRelatedWork W2044225567 @default.
- W3048497507 hasRelatedWork W2095723734 @default.
- W3048497507 hasRelatedWork W2105918531 @default.
- W3048497507 hasRelatedWork W2169163411 @default.
- W3048497507 hasRelatedWork W2182010643 @default.
- W3048497507 hasRelatedWork W2287860957 @default.
- W3048497507 hasRelatedWork W2288740352 @default.