Novel engine trained on human-played games

Collapse
X
 
  • Time
  • Show
Clear All
new posts
  • Kerry Liles
    • 2008-05-30
    • 3071

    #1

    Novel engine trained on human-played games

    An article from Microsoft Research:



    As AI-powered chess engines get smarter and interest in the game increases, Microsoft researchers and their collaborators at the University of Toronto and Cornell University seek to better match AI to different human skill levels and playing styles. They introduce Maia. Trained on games played by people, Maia matches human play better than any other chess engine, providing promising insight into the future of human-AI collaboration and learning.
    ...Mike Pence: the Lord of the fly.
  • Vadim Tsypin
    • 2013-12-31
    • 690

    #2
    Importantly, every version of Maia uniquely captures a specific human skill level since every curve achieves its maximum accuracy at a different human rating. Even Maia 1100 achieves over 50% accuracy in predicting 1100-rated moves, and it’s much better at predicting 1100-rated players than 1900-rated players!

    This means something deep about chess: there is such a thing as “1100-rated style.” And furthermore, it can be captured by a machine learning model.
    Thank you very much Kerry! This is fascinating.

    Comment

    • Hans Jung
      • 2008-06-01
      • 21137

      #3
      Noteworthy.

      Comment

      • Garland Best
        • 2008-05-30
        • 722

        #4
        Originally posted by Vadim Tsypin View Post
        This means something deep about chess: there is such a thing as “1100-rated style.” And furthermore, it can be captured by a machine learning model.
        Terrific. So now it will become even harder to catch cheaters. As a 1600 rated player, I could use a 1900 rated human-like program and clean up in a sectional tournament without getting caught. Ain't technology wonderful!

        Comment

        Working...