Online Bandit Learning against an Adaptive Adversary: from Regret to Policy Regret.

Ofer Dekel, Ambuj Tewari, Raman Arora

Browse the full ICML paper archive.