Darkforest

This article is about Facebook's Go-playing computer program. For other uses, see Dark Forest.

Darkforest is a computer go program developed by Meta Platforms, based on deep learning techniques using a convolutional neural network. Its updated version Darkfores2 combines the techniques of its predecessor with Monte Carlo tree search.[1][2] The MCTS effectively takes tree search methods commonly seen in computer chess programs and randomizes them.[3] With the update, the system is known as Darkfmcts3.[4]

Darkforest is of similar strength to programs like CrazyStone and Zen.[5] It has been tested against a professional human player at the 2016 UEC cup. Google's AlphaGo program won against a professional player in October 2015 using a similar combination of techniques.[6]

Darkforest is named after Liu Cixin's science fiction novel The Dark Forest.[7]

  1. ^ Tian, Yuandong; Zhu, Yan (2015). "Better Computer Go Player with Neural Network and Long-term Prediction". arXiv:1511.06410v1 [cs.LG].
  2. ^ "How Facebook's AI Researchers Built a Game-Changing Go Engine". MIT Technology Review. December 4, 2015. Retrieved 2016-02-03.
  3. ^ "Facebook AI Go Player Gets Smarter With Neural Network And Long-Term Prediction To Master World's Hardest Game". Tech Times. 2016-01-28. Retrieved 2016-04-24.
  4. ^ "Facebook's artificially intelligent Go player is getting smarter". VentureBeat. 27 January 2016. Retrieved 2016-04-24.
  5. ^ "Strachey Lecture - Dr Demis Hassabis by University of Oxford Live".
  6. ^ "No Go: Facebook fails to spoil Google's big AI day". The Guardian. 2016-01-28. ISSN 0261-3077. Retrieved 2016-02-01.
  7. ^ "FB围棋项目负责人谈人机大战" [FB Go Project Manager Discusses Man vs Machine Showdown] (in Chinese). Tencent. 2016-03-01.