Model

• PolicyNet: Neural network estimating move probabilities (policy) for a given board state.
• ValueNet: Neural network estimating the value (expected outcome) of a board state.
• DualNetGoBot: Combines PolicyNet and ValueNet for reinforcement learning; selects moves using the policy net and evaluates positions with the value net.