AI在Stratego中击败史上最强人类选手，训练成本不到8000美元
原标题：AI beats Stratego's greatest player, ending one of the last human strongholds in board games
AI 摘要
卡内基梅隆大学、纽约大学、斯坦福大学和麻省理工学院的研究团队开发了名为Ataraxos的Stratego AI，在Nature论文中报告其首次在该棋盘游戏达到超人水平。在20局正式系列赛中以15胜1负4平击败史上最成功选手Pim Niemeijer。训练仅用16块Nvidia H100 GPU一周加4块GPU四天，成本不到8000美元，远低于DeepMind DeepNash估计的300万至450万美元。
正文节选
AI beats Stratego's greatest player, ending one of the last human strongholds in board games A team from Carnegie Mellon, NYU, Stanford, and MIT has built an AI that decisively beats the most successful Stratego player of all time. Training reportedly cost less than $8,000. In a paper published in Nature, researchers from Carnegie Mellon University, New York University, Stanford University, and MIT introduce Ataraxos, a Stratego AI. To the authors' knowledge, it's the first AI to reach superhuma