The History of AI

From Atari to Go

In four years, DeepMind goes from learning video games from pixels to mastering Go without human games.

  1. DQN (December 19, 2013)

    A convolutional network learns to play Atari video games from nothing but the pixels and the score.

  2. AlphaGo beats Fan Hui (October 5, 2015)

    The same lab combines deep networks, reinforcement learning, and tree search, and AlphaGo beats Fan Hui, the European Go champion, 5-0.

  3. AlphaGo beats Lee Sedol (March 9, 2016)

    An improved version of AlphaGo beats Lee Sedol 4-1 in Seoul.

  4. AlphaGo Zero and AlphaZero (October 19, 2017)

    AlphaGo Zero does without human games and beats the version that defeated Lee Sedol 100-0. AlphaZero applies the same method to chess and shogi.

Open in the interactive timeline