Multi-Armed Bandits with Betting

We study an extension to the stochastic multiarmed bandit problem where the learner has a budget ofK “coins” it can use in each round. The learner can use the coins to play multiple arms in each r