Text this: Bayesian Reinforcement Learning With Limited Cognitive Load