Stochastic Regret Minimization via Thompson Sampling

The Thompson Sampling (TS) policy is a widely implemented algorithm for the stochastic multi-armed bandit (MAB) problem. Given a prior distribution over possible parameter settings of the underlying r