Imperfect-Information Games
Inside imperfect-information games: how AI makes decisions when it cannot see the opponent's hand — tracking algorithmic breakthroughs and match results from Stratego, poker, mahjong and beyond.
Most real contests run on hidden information: you cannot see the opponent's cards or positions, and imperfect-information games study optimal decisions in that setting. From DeepMind's DeepNash to Ataraxos, published in Nature, AI has surpassed top Stratego players, while poker, mahjong, negotiation and cybersecurity become proving grounds. This topic records how these systems calculate risk, see through bluffs, and enter the real world.
