OnlineMulti-ArmedBanditswithAdaptiveInference

During online decision making in multi-armed bandits, one needs to conduct inference on the true mean reward of each arm based on data collected so far at each step.