On the Pareto Frontier of Regret Minimization and Best Arm Identification in Stochastic Bandits

Open in new window