Stratego

Researchers have built an AI that beat the best Stratego player in history, and did so cheaply. Because most pieces in Stratego are hidden, the system added a second neural network that guesses the identity of hidden pieces, which proved key to handling the game's imperfect information. The result is drawing attention as a milestone for AI in hidden-information games.

Researchers report a new game-playing AI that has beaten human players at Stratego, a board game of hidden information long considered a major challenge for artificial intelligence. The system is said to rely on highly efficient self-play reinforcement learning and can also play Barrage Stratego, Hanabi, and Dou dizhu, suggesting broad capabilities across imperfect-information games.

An AI system called Ataraxos has beaten Pim Niemeijer, widely regarded as the best Stratego player ever, according to coverage referencing publication in Nature. The result is being shared in tech circles as a notable milestone in game-playing AI, since Stratego involves hidden information and bluffing, making it harder for machines than games like chess or Go.

An AI system has become the new champion at Stratego, the classic board game long considered a major challenge for artificial intelligence because players cannot see each other's pieces. The result, reported by MIT News, marks a step forward in AI's ability to handle imperfect-information games, beyond chess and Go.