Лекция 6. Actor-Critic и Proximal Policy Optimization