A general reinforcement learning algorithm that masters chess, shogi, and Go through self-play
AlphaZero teaches itself to play three different board games and beats state-of-the-art programs in each.
science.sciencemag.org