MVA-2021 / reinforcement_learning / hw3_exploration