We formulate the novel class of contextual games, a type of repeated games driven by contextual information at each round . By means of kernel-based regularity assumptions, we model the correlation between different contexts and outcomes . We propose a novel online algorithm that exploits suchcorrelations to minimize the contextual regret of individual players . We show that c-CCEs andoptimal welfare can be approached whenever players’ contextual regrets vanish . We empirically validate our results in a traffic routing experiment, where our algorithm leads to better performance and higher welfare compared tobaselines that do not exploit the available contextual information .

Author(s) : Pier Giuseppe Sessa, Ilija Bogunovic, Andreas Krause, Maryam Kamgarpour

Links : PDF - Abstract

Code :

Keywords : contextual - games - information - players - welfare -

Leave a Reply

Your email address will not be published. Required fields are marked *