PhD Seminar Talk: "The Necessary Optimality Conditions for Deep Reinforcement Learning- Challenges and Benefits"