Theory, with a particular emphasis on sequential decision-making in uncertain environments. Sequential hypothesis testing and multi-armed bandit frameworks. Stochastic approximation techniques to advance theoretical understanding of reinforcement learning