Accueil > Manifestations > Séminaires du CMLA
Recherche - Valorisation
Séminaire du CMLA : New Results at the Crossroads of Convexity, Learning and Information Theory
le 30 juin 2016
14h30 - 15h30
Séminaire du CMLA animé par Sébastien Bubeck (Microsoft Research, Princeton)
I will present
three new results at the intersection of convex geometry, learning and information theory :
(no special background in these topics or in optimization will be assumed) (i) the
Cramer transform of the uniform measure on a convex body is a universal self-concordant barrier;
(ii) projected gradient descent with
Gaussian noise allows to sample from a log-concave measure in polynomial time;
and (iii)
Thompson sampling combined with a multi-scale exploration solves the
Bayesian convex bandit problem.
- Type :
- Séminaires - conférences
- Lieu(x) :
- Campus de Cachan
Salle des conférences, Pavillon des Jardins