Policy Gradient Methods Tasks | State-of-the-Art