David Stern
Kernel Topic Models.
Mon, 10/24/2011 - 15:01 — SomNambulAuthors: David Stern, Philipp Hennig, Ralf Herbrich, Thore Graepel
Subjects: Learning
AbstractLatent Dirichlet Allocation models discrete data as a mixture of discrete
distributions, using Dirichlet beliefs over the mixture weights. We study a
variation of this concept, in which the documents' mixture weight beliefs are
replaced with squashed Gaussian distributions. This allows documents to be
associated with elements of a Hilbert space, admitting kernel topic models
(KTM), modelling temporal, spatial, hierarchical, social and other structure
between documents. The main challenge is efficient approximate inference on the
latent Gaussian.Helices on del Pezzo surfaces and tilting Calabi-Yau algebras.
Thu, 09/10/2009 - 10:12 — SomNambulAuthors: Tom Bridgeland, David Stern
Subjects: Rings and Algebras
AbstractWe study tilting for a class of Calabi-Yau algebras associated to helices on
Fano varieties. We do this by relating the tilting operation to mutations of
exceptional collections. For helices on del Pezzo surfaces the algebras are of
dimension three, and using an argument of Herzog, together with results of
Kuleshov and Orlov, we obtain a complete description of the tilting process in
terms of quiver mutations.Helices on del Pezzo surfaces and tilting Calabi-Yau algebras.
Thu, 09/10/2009 - 10:12 — SomNambulAuthors: Tom Bridgeland, David Stern
Subjects: Rings and Algebras
AbstractWe study tilting for a class of Calabi-Yau algebras associated to helices on
Fano varieties. We do this by relating the tilting operation to mutations of
exceptional collections. For helices on del Pezzo surfaces the algebras are of
dimension three, and using an argument of Herzog, together with results of
Kuleshov and Orlov, we obtain a complete description of the tilting process in
terms of quiver mutations.
User login
