Dec 2017
A general reinforcement-learning approach learns chess, shogi, and Go from self-play.
A shared algorithm replaces some game-specific engineering.
Explore on Humanity Unfolds
