Αποστολή με SMS: Adaptive representations for reinforcement learning