Researchers from MIT, Carnegie Mellon, NYU and Stanford have developed Ataraxos, an AI system that surpassed top human players in Stratego, a strategy game built around hidden information.

It combines self-play reinforcement learning with real-time planning that estimates the most likely hidden state of the game while requiring far less computation than earlier approaches. Ataraxos posted a 39-2 record against top human competitors, and the researchers say the underlying method could eventually support decisions in areas such as negotiations and cybersecurity.