Improving Monte Carlo Tree Search Policies in StarCraft via Probabilistic Models Learned from Replay Data

Open in new window