04. Reinforcement Learning - AlphaGoECM/AlphaGo GitHub Wiki

At this moment of the training, we obtained a policy network, which learned to predict experts' moves from hundred of thousands of couples (state, action), through supervised learning.

Now, the moment is to reinforce the policy network and create the value network