In that approach, the algorithm finds a winning strategy by applying a set of tactical moves that can’t be exploited by the opponent. Now a DeepMind team led by Julien Perolat, Bart De Vylder, and Karl Tuyls has developed an algorithm called DeepNash that plays Stratego at the level of a human expert.ĭeepNash plays at a highly competitive level by finding a Nash equilibrium. The research laboratory DeepMind Technologies became famous in 2016 when its AlphaGo algorithm beat Go world champion Lee Sedol in a five-game match. (Only during an interaction between pieces do their ranks become known.) For an AI algorithm to win, it must make a series of long-term strategic moves and analyze a staggering 10 60 times as many starting arrangements as a two-player game of Texas Hold ’em. Two players each control 40 pieces: A piece can capture one of lower rank, but the specific ranks of the opponent’s pieces are unknown. As in capture the flag, each player guards their flag and tries to capture their opponent’s. That number, however, is but a fraction of the 10 535 possible states for the board game Stratego. Despite the added complexity of the game compared with perfect-information games like chess, in 2015 artificial intelligence (AI) researchers designed a game-winning strategy for Texas Hold ’em, a variation of poker with 10 164 possible game states. Because each player has a set of starting cards that others can’t see, a player can bluff. Credit: zizou man, Wikimedia Commons, CC BY 2.0Ī game like poker is one with imperfect information. An opening arrangement of the imperfect-information game Stratego.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |