Multitask Learning with No Regret: from Improved Confidence Bounds to Active Learning

Open in new window