Learning-How-To-Act Design and Analysis Of Experiments Observational Causal Inference Reinforcement Learning - Multi Arm Bandit Algorithms