PAC-Bayesian aggregation and multi-armed bandits

Jean-Yves Audibert (1, 2, 3)

1 imagine [Marne-la-Vallée]
2 WILLOW - Models of visual object recognition and scene understanding
3 LIGM - Laboratoire d'Informatique Gaspard-Monge

This habilitation thesis presents several contributions to (1) the PAC-Bayesian analysis of statistical learning, (2) the three aggregation problems: given d functions, how to predict as well as (i) the best of these d functions (model selection type aggregation), (ii) the best convex combination of these d functions, (iii) the best linear combination of these d functions, (3) the multi-armed bandit problems.

Mots clés: statistical learning; aggregation problems

Domaines: Statistiques [math.ST]; Théorie [stat.TH]

Jean-Yves Audibert. PAC-Bayesian aggregation and multi-armed bandits. Statistics [math.ST]. Université Paris-Est, 2010. ⟨tel-00536084⟩