AlphaGo Prepares for Its Match Against Lee Sedol

In January 2016, DeepMind published the Nature paper describing AlphaGo and announced that the system would face Lee Sedol, one of the world’s strongest Go players. The paper was significant because Go had long resisted the brute-force search methods that worked well in games such as chess.

AlphaGo combined several techniques. A policy network predicted promising moves, a value network estimated who was likely to win from a given board position, and Monte Carlo tree search explored selected continuations instead of trying to examine every legal move. Earlier versions were trained first from expert human games and then improved through self-play.

The system had already defeated European champion Fan Hui 5–0 in a formal match. The planned Lee Sedol contest would test whether the same architecture could handle a player operating at the very top professional level.