Act as a world-class AI researcher. I want to create a new Reinforcement Learning (RL) method. RL is when computers learn from rewards. This method must be as fast as simple math but as strong as deep learning. Do not use geometry. 1) First, explain your plan for this new math. 2) List the steps you will take to prove it works. 3) Once I agree, write the new formulas. Use clear names for all parts...