Massimiliano Amarante – Université de Montréal, Canada
As these concepts will be freely used in the second portion of the talk, this part quickly reviews some of the main ideas in the theory of decision making under uncertainty such as Expected Utility (EU), Maxmin EU, Choquet EU, the Invariant bi-separable model and the Variational model. We will discuss the relation of these ideas with problems like one-shot learning or learning with corrupted data as well as applications to Robust Statistics, fuzzy systems, climate change and finance. As a bridge to the second part, we will talk about the difficulties encountered in extending non-Bayesian models to a dynamic setting.
Partie 2 à 11h15
Campus de l'Université de Montréal
2920, chemin de la Tour
Montréal Québec H3T 1J4