The State-of-the-Art Mandate

Reinforcement learning arrives as an instruction, not a question and the supervision you actually hold, not the technique you read about, is what decides.