Reinforcement learning, supervised learning, diffusion models