Exploring the Reinforcement Learning Techniques in AlphaGo