We consider multiarmed bandit problems where the expected reward isunimodal over a partially ordered set of arms. In particular, thearms may belong to a continuous interval or correspond to verticesin a graph. We present efficient algorithms to minimize the regretin these bandit problems and also to detect abrupt changes in thereward distributions. The unimodality assumption has an importantadvantage: we can determine if a given arm is optimal by samplingthe possible directions around it. This property allows us toquickly find the optimal arm in a graph and detect changes. Notably,our method […]
In this paper, we consider the global wellposedness of the 3-D incompressible anisotropic Navier-Stokes equations with initial data in the critical Besov-Sobolev type spaces. This is partially joint work with J. Y. Chemin, Guilong GUI and M. Paicu.
I shall talk about some regularity results for the ultraparabolic equation, in particular, the C^{alpha}$ regularity of weak solutions. The problem arises from the Prandtl's boundary layer system under the Crocco transformation. I shall also report some recent results on the backward uniqueness of ultraparabolic equations.
Equations de Navier-Stokes
On étend la définition des espaces homogènes sphériques et de leurs plongements au cas d'un corps quelconque. On montre qu'à un plongement d'un espace homogène sphérique fixé X, on peut associer un éventail colorié stable par le groupe de Galois. On présente des exemples où cette correpondance est parfaite.
Soit X une variété projective lisse sur in corps de nombres, fibrée au dessus d'une courbe C, à fibres géométriquement intègres. En supposant que les fibres d'un sous-ensemble hilbertien généralisé satisfont le principe de Hasse (resp. l'approximation faible) et la finitude du groupe de Tate-Schafarevitch de la jacobienne de C), on montre que l'obstruction de Brauer-Manin provenant de la courbe d'en bas est la seule au principe de Hasse (resp. à l'approximation faible) pour les zéros-cycles de degré 1 sur X.
Exposé en français mais transparents en anglais I will present recent works on least-squares regression using randomly generated subspaces.In this approach, the regression function is the empirical risk minimizer in a low dimensional randomly generated subspace of a high (possibly infinite) dimensional function space. This approach can be seen as an alternative to usual penalization techniques. Approximation error and excess risk bounds are derived and the issue of numerical complexity will be discussed.This is joint work with Odalric Maillard and is described in the following papers: - Compressed Least-Squares […]