Plusieurs problèmes importants issus de l'apprentissage statistique et de la science des données impliquent des objectifs d'optimisation à très haute dimension qui vont au delà des hypothèses standard de régularité de Lipschitz. L'absence de régularité de Lipschitz - lissage ou continuité - pose des défis importants à l'analyse de convergence de la plupart des algorithmes existants d'optimisation et, dans de nombreux cas, elle nécessite de nouveaux outils analytiques et algorithmiques pour être traitée efficacement. Dans cette thèse, nous visons à combler partiellement cette lacune par la conception et l'analyse de nouvelles méthodes universelles du premier ordre pour deux cadres d'optimisation généraux : (a) l'optimisation convexe en ligne...
This paper deals with two kinds of the one-dimensional global optimization problem over a closed fin...
Global optimization methods based on Lipschitz bounds have been analyzed and applied widely to solve...
In this thesis, Direct (DIviding RECTangles) type algorithms based on Lipschitz objective function m...
Several important problems in learning theory and data science involve high-dimensional optimization...
International audienceWe propose a new family of adaptive first-order methods for a class of convex ...
Ce travail de thèse s’intéresse au problème d’optimisation séquentielle d’une fonction inconnue défi...
We consider a family of function classes which allow functions with several minima and which deman...
Recently there were proposed some innovative convex optimization concepts, namely, relative smoothne...
AbstractWe consider the global optimization problem for d-variate Lipschitz functions which, in a ce...
In this work, we present a new deterministic partition-based Global Optimization (GO) algorithm that...
International audienceWe present a new family of min-max optimization algorithms that automatically ...
The paper discusses how the used norm and corresponding Lipschitz constant influence the speed of al...
Many problems in economy may be formulated as global optimization problems. Most numerically promisi...
The global optimisation problem has received much attention in the past twenty-five years. The probl...
This work addresses the sequential optimization of an unknown and potentially nonconvex function ove...
This paper deals with two kinds of the one-dimensional global optimization problem over a closed fin...
Global optimization methods based on Lipschitz bounds have been analyzed and applied widely to solve...
In this thesis, Direct (DIviding RECTangles) type algorithms based on Lipschitz objective function m...
Several important problems in learning theory and data science involve high-dimensional optimization...
International audienceWe propose a new family of adaptive first-order methods for a class of convex ...
Ce travail de thèse s’intéresse au problème d’optimisation séquentielle d’une fonction inconnue défi...
We consider a family of function classes which allow functions with several minima and which deman...
Recently there were proposed some innovative convex optimization concepts, namely, relative smoothne...
AbstractWe consider the global optimization problem for d-variate Lipschitz functions which, in a ce...
In this work, we present a new deterministic partition-based Global Optimization (GO) algorithm that...
International audienceWe present a new family of min-max optimization algorithms that automatically ...
The paper discusses how the used norm and corresponding Lipschitz constant influence the speed of al...
Many problems in economy may be formulated as global optimization problems. Most numerically promisi...
The global optimisation problem has received much attention in the past twenty-five years. The probl...
This work addresses the sequential optimization of an unknown and potentially nonconvex function ove...
This paper deals with two kinds of the one-dimensional global optimization problem over a closed fin...
Global optimization methods based on Lipschitz bounds have been analyzed and applied widely to solve...
In this thesis, Direct (DIviding RECTangles) type algorithms based on Lipschitz objective function m...