Learning the distribution with largest mean: two bandit frameworksEmilie Kaufmann and Aurélien GarivierESAIM: Procs, 60 (2017) 114-131DOI: https://doi.org/10.1051/proc/201760114
