Reinforcement learning/3. Frontiers 0