Matches in SemOpenAlex for { <https://semopenalex.org/work/W2113111165> ?p ?o ?g. }
- W2113111165 endingPage "1059" @default.
- W2113111165 startingPage "1051" @default.
- W2113111165 abstract "The problem of approximating a given probability distribution using a simpler distribution plays an important role in several areas of machine learning, for example variational inference and classification. Within this context, we consider the task of learning a mixture of tree distributions. Although mixtures of trees can be learned by minimizing the KL-divergence using an EM algorithm, its success depends heavily on the initialization. We propose an efficient strategy for obtaining a good initial set of trees that attempts to cover the entire observed distribution by minimizing the α-divergence with α = ∞. We formulate the problem using the fractional covering framework and present a convergent sequential algorithm that only relies on solving a convex program at each iteration. Compared to previous methods, our approach results in a significantly smaller mixture of trees that provides similar or better accuracies. We demonstrate the usefulness of our approach by learning pictorial structures for face recognition." @default.
- W2113111165 created "2016-06-24" @default.
- W2113111165 creator A5022356850 @default.
- W2113111165 creator A5051658526 @default.
- W2113111165 date "2009-12-07" @default.
- W2113111165 modified "2023-09-24" @default.
- W2113111165 title "Learning a Small Mixture of Trees" @default.
- W2113111165 cites W1524704912 @default.
- W2113111165 cites W1530235965 @default.
- W2113111165 cites W1555384764 @default.
- W2113111165 cites W1817561967 @default.
- W2113111165 cites W1988790447 @default.
- W2113111165 cites W2024046085 @default.
- W2113111165 cites W2045798786 @default.
- W2113111165 cites W2058168712 @default.
- W2113111165 cites W2061130579 @default.
- W2113111165 cites W2095979710 @default.
- W2113111165 cites W2096222410 @default.
- W2113111165 cites W2098199022 @default.
- W2113111165 cites W2103013841 @default.
- W2113111165 cites W2113656553 @default.
- W2113111165 cites W2121805075 @default.
- W2113111165 cites W2123761428 @default.
- W2113111165 cites W2126559756 @default.
- W2113111165 cites W2128547631 @default.
- W2113111165 cites W2134422938 @default.
- W2113111165 cites W2135094946 @default.
- W2113111165 cites W2159080219 @default.
- W2113111165 cites W2163166770 @default.
- W2113111165 cites W2163664777 @default.
- W2113111165 cites W2164918853 @default.
- W2113111165 cites W2168996682 @default.
- W2113111165 cites W2296319761 @default.
- W2113111165 cites W568057840 @default.
- W2113111165 hasPublicationYear "2009" @default.
- W2113111165 type Work @default.
- W2113111165 sameAs 2113111165 @default.
- W2113111165 citedByCount "4" @default.
- W2113111165 countsByYear W21131111652012 @default.
- W2113111165 countsByYear W21131111652014 @default.
- W2113111165 countsByYear W21131111652016 @default.
- W2113111165 countsByYear W21131111652021 @default.
- W2113111165 crossrefType "proceedings-article" @default.
- W2113111165 hasAuthorship W2113111165A5022356850 @default.
- W2113111165 hasAuthorship W2113111165A5051658526 @default.
- W2113111165 hasConcept C112680207 @default.
- W2113111165 hasConcept C113174947 @default.
- W2113111165 hasConcept C11413529 @default.
- W2113111165 hasConcept C114466953 @default.
- W2113111165 hasConcept C114614502 @default.
- W2113111165 hasConcept C119857082 @default.
- W2113111165 hasConcept C126255220 @default.
- W2113111165 hasConcept C138885662 @default.
- W2113111165 hasConcept C144024400 @default.
- W2113111165 hasConcept C151730666 @default.
- W2113111165 hasConcept C153180895 @default.
- W2113111165 hasConcept C154945302 @default.
- W2113111165 hasConcept C177264268 @default.
- W2113111165 hasConcept C199360897 @default.
- W2113111165 hasConcept C207390915 @default.
- W2113111165 hasConcept C2524010 @default.
- W2113111165 hasConcept C2776214188 @default.
- W2113111165 hasConcept C2779304628 @default.
- W2113111165 hasConcept C2779343474 @default.
- W2113111165 hasConcept C33923547 @default.
- W2113111165 hasConcept C36289849 @default.
- W2113111165 hasConcept C41008148 @default.
- W2113111165 hasConcept C41895202 @default.
- W2113111165 hasConcept C86803240 @default.
- W2113111165 hasConceptScore W2113111165C112680207 @default.
- W2113111165 hasConceptScore W2113111165C113174947 @default.
- W2113111165 hasConceptScore W2113111165C11413529 @default.
- W2113111165 hasConceptScore W2113111165C114466953 @default.
- W2113111165 hasConceptScore W2113111165C114614502 @default.
- W2113111165 hasConceptScore W2113111165C119857082 @default.
- W2113111165 hasConceptScore W2113111165C126255220 @default.
- W2113111165 hasConceptScore W2113111165C138885662 @default.
- W2113111165 hasConceptScore W2113111165C144024400 @default.
- W2113111165 hasConceptScore W2113111165C151730666 @default.
- W2113111165 hasConceptScore W2113111165C153180895 @default.
- W2113111165 hasConceptScore W2113111165C154945302 @default.
- W2113111165 hasConceptScore W2113111165C177264268 @default.
- W2113111165 hasConceptScore W2113111165C199360897 @default.
- W2113111165 hasConceptScore W2113111165C207390915 @default.
- W2113111165 hasConceptScore W2113111165C2524010 @default.
- W2113111165 hasConceptScore W2113111165C2776214188 @default.
- W2113111165 hasConceptScore W2113111165C2779304628 @default.
- W2113111165 hasConceptScore W2113111165C2779343474 @default.
- W2113111165 hasConceptScore W2113111165C33923547 @default.
- W2113111165 hasConceptScore W2113111165C36289849 @default.
- W2113111165 hasConceptScore W2113111165C41008148 @default.
- W2113111165 hasConceptScore W2113111165C41895202 @default.
- W2113111165 hasConceptScore W2113111165C86803240 @default.
- W2113111165 hasLocation W21131111651 @default.
- W2113111165 hasOpenAccess W2113111165 @default.
- W2113111165 hasPrimaryLocation W21131111651 @default.
- W2113111165 hasRelatedWork W2097575529 @default.
- W2113111165 hasRelatedWork W2103013841 @default.