Concentration inequalities for functions of independent random variables

schedule le lundi 11 mars 2019 de 17h00 à 18h00

Organisé par : A. Lefebvre, N. Meyer, O. Safsafi, T. Touati

Intervenant : Antoine Marchina (Université Paris-Est)
Lieu : Jussieu, salle de séminaire, couloir 16-26, salle 209

Sujet : Probabilités

Résumé :

This talk deals with concentration properties around the mean of separately convex functions of independent random variables which are not necessarily bounded. I present the method used which is based on martingale techniques associated with Pinelis type comparison inequalities. To this end, I establish comparison inequalities for stochastically dominated random variables (on the left and on the right) that generalize a classical inequality on bounded random variables originally due to Hoeffding.