Matches in SemOpenAlex for { <https://semopenalex.org/work/W2949759968> ?p ?o ?g. }
- W2949759968 abstract "Estimating and optimizing Mutual Information (MI) is core to many problems in machine learning; however, bounding MI in high dimensions is challenging. To establish tractable and scalable objectives, recent work has turned to variational bounds parameterized by neural networks, but the relationships and tradeoffs between these bounds remains unclear. In this work, we unify these recent developments in a single framework. We find that the existing variational lower bounds degrade when the MI is large, exhibiting either high bias or high variance. To address this problem, we introduce a continuum of lower bounds that encompasses previous bounds and flexibly trades off bias and variance. On high-dimensional, controlled problems, we empirically characterize the bias and variance of the bounds and their gradients and demonstrate the effectiveness of our new bounds for estimation and representation learning." @default.
- W2949759968 created "2019-06-27" @default.
- W2949759968 creator A5004918186 @default.
- W2949759968 creator A5036325834 @default.
- W2949759968 creator A5039776447 @default.
- W2949759968 creator A5048032272 @default.
- W2949759968 creator A5049332971 @default.
- W2949759968 date "2019-05-16" @default.
- W2949759968 modified "2023-10-01" @default.
- W2949759968 title "On Variational Bounds of Mutual Information" @default.
- W2949759968 cites W115285041 @default.
- W2949759968 cites W1508769776 @default.
- W2949759968 cites W1686946872 @default.
- W2949759968 cites W1959608418 @default.
- W2949759968 cites W2092939357 @default.
- W2949759968 cites W2097482982 @default.
- W2949759968 cites W2097636360 @default.
- W2949759968 cites W2101484286 @default.
- W2949759968 cites W2108384452 @default.
- W2949759968 cites W2114771311 @default.
- W2949759968 cites W2136144249 @default.
- W2949759968 cites W2148349024 @default.
- W2949759968 cites W2165700458 @default.
- W2949759968 cites W2166944917 @default.
- W2949759968 cites W2321533354 @default.
- W2949759968 cites W2439880944 @default.
- W2949759968 cites W2557579533 @default.
- W2949759968 cites W2562420335 @default.
- W2949759968 cites W2613509922 @default.
- W2949759968 cites W2785519580 @default.
- W2949759968 cites W2785885194 @default.
- W2949759968 cites W2787273002 @default.
- W2949759968 cites W2789776893 @default.
- W2949759968 cites W2803439868 @default.
- W2949759968 cites W2803832867 @default.
- W2949759968 cites W2804233805 @default.
- W2949759968 cites W2807692250 @default.
- W2949759968 cites W2842511635 @default.
- W2949759968 cites W2892806280 @default.
- W2949759968 cites W2899477140 @default.
- W2949759968 cites W2900148439 @default.
- W2949759968 cites W2905204718 @default.
- W2949759968 cites W2908689144 @default.
- W2949759968 cites W2950365520 @default.
- W2949759968 cites W2950662112 @default.
- W2949759968 cites W2950797609 @default.
- W2949759968 cites W2951025442 @default.
- W2949759968 cites W2951873722 @default.
- W2949759968 cites W2952673310 @default.
- W2949759968 cites W2962679290 @default.
- W2949759968 cites W2962738009 @default.
- W2949759968 cites W2962897886 @default.
- W2949759968 cites W2963366547 @default.
- W2949759968 cites W2963636093 @default.
- W2949759968 cites W2963800509 @default.
- W2949759968 cites W2963926704 @default.
- W2949759968 cites W2964184826 @default.
- W2949759968 cites W343636949 @default.
- W2949759968 hasPublicationYear "2019" @default.
- W2949759968 type Work @default.
- W2949759968 sameAs 2949759968 @default.
- W2949759968 citedByCount "106" @default.
- W2949759968 countsByYear W29497599682019 @default.
- W2949759968 countsByYear W29497599682020 @default.
- W2949759968 countsByYear W29497599682021 @default.
- W2949759968 countsByYear W29497599682022 @default.
- W2949759968 crossrefType "posted-content" @default.
- W2949759968 hasAuthorship W2949759968A5004918186 @default.
- W2949759968 hasAuthorship W2949759968A5036325834 @default.
- W2949759968 hasAuthorship W2949759968A5039776447 @default.
- W2949759968 hasAuthorship W2949759968A5048032272 @default.
- W2949759968 hasAuthorship W2949759968A5049332971 @default.
- W2949759968 hasConcept C11413529 @default.
- W2949759968 hasConcept C121955636 @default.
- W2949759968 hasConcept C126255220 @default.
- W2949759968 hasConcept C134306372 @default.
- W2949759968 hasConcept C144133560 @default.
- W2949759968 hasConcept C152139883 @default.
- W2949759968 hasConcept C154945302 @default.
- W2949759968 hasConcept C165464430 @default.
- W2949759968 hasConcept C17744445 @default.
- W2949759968 hasConcept C196083921 @default.
- W2949759968 hasConcept C199539241 @default.
- W2949759968 hasConcept C2776359362 @default.
- W2949759968 hasConcept C33923547 @default.
- W2949759968 hasConcept C41008148 @default.
- W2949759968 hasConcept C48044578 @default.
- W2949759968 hasConcept C63584917 @default.
- W2949759968 hasConcept C77088390 @default.
- W2949759968 hasConcept C77553402 @default.
- W2949759968 hasConcept C94625758 @default.
- W2949759968 hasConceptScore W2949759968C11413529 @default.
- W2949759968 hasConceptScore W2949759968C121955636 @default.
- W2949759968 hasConceptScore W2949759968C126255220 @default.
- W2949759968 hasConceptScore W2949759968C134306372 @default.
- W2949759968 hasConceptScore W2949759968C144133560 @default.
- W2949759968 hasConceptScore W2949759968C152139883 @default.
- W2949759968 hasConceptScore W2949759968C154945302 @default.
- W2949759968 hasConceptScore W2949759968C165464430 @default.
- W2949759968 hasConceptScore W2949759968C17744445 @default.