UCB algorithms for multi-armed bandits: Precise regret and adaptive inference

Re-assign community