Panayotis Mertikopoulos

interaction seminar

Panayotis Mertikopoulos

CNRS
No-regret learning in games
Lieu

IBD Salle 21

Îlot Bernard du Bois - Salle 21

AMU - AMSE
5-9 boulevard Maurice Bourdet
13001 Marseille

Date(s)
Jeudi 30 novembre 2017| 12:00 - 13:15
Contact(s)

Ugo Bolletta : ugo.bolletta[at]univ-amu.fr
Mathieu Faure : mathieu.faure[at]univ-amu.fr

Résumé

In many cases of practical interest, the players of a repeated game may not know the structure of the game being played or even that they are involved in a game - simply think of commuters driving to work every day, ignorant of the number of vehicles at each part of the road. In such cases, it is often assumed that players follow a "no-regret" procedure, i.e. an updating policy that probably minimizes each player's individual regret against any possible play of their opponents.

This talk focuses on the following question: does the sequence of play induced by no-regret learning converge to an equilibrium of the underlying stage game? I will present some recent contributions to this question (in both finite and continuous games), and I will discuss the impact of the feedback available to the players.