Stochastic policy gradient

1. State-Action-Reward-State-Action: AI (Brace For These Hidden GPT Dangers)
2. Soft Actor-Critic: AI (Brace For These Hidden GPT Dangers)