Reinforcement learning In just a few hours, DeepMind's AI AlphaZero learn to master both chess and shogi, beating top chess software Stockfish and Elmo without domain knowledge. Sources: Critical Hit