RESEARCH · RESEARCH · #1280
Ataraxos AI from MIT-led team achieves superhuman performance in Stratego
Researchers from MIT, Carnegie Mellon, NYU and Stanford developed Ataraxos, an AI trained with self-play and efficient algorithms that defeated top-ranked human Stratego players by a large margin and outperformed prior models while requiring far less compute; the work is reported in Nature. The system also generalizes to other imperfect-information strategic games and could inform decision-making in real-world hidden-information problems.
KEY POINTS
- Researchers from MIT, Carnegie Mellon, NYU and Stanford developed Ataraxos, an AI trained with self-play and efficient algorithms that defeated top-ranked human Stratego players by a large margin and outperformed prior models while requiring far less compute; the work is reported in Nature.
- The system also generalizes to other imperfect-information strategic games and could inform decision-making in real-world hidden-information problems.
- Demonstrates a more compute-efficient, general approach to solving large imperfect-information problems, with implications for negotiation, cybersecurity and military planning.
WHY IT MATTERS
Demonstrates a more compute-efficient, general approach to solving large imperfect-information problems, with implications for negotiation, cybersecurity and military planning.