Alpha Zero
A successor to AlphaGo Zero, Alpha Zero was an attempt at generalizing the algorithm so that it could be used to play other board games as well. Alpha Zero was trained to play chess, shogi (a Japanese game similar to chess), and Go and achieved performance on par with the existing AIs of the respective games. After 34 hours of training, Alpha Zero for Go was able to defeat AlphaGo Zero, which was trained over 3 days, with a score of 60:40. This led to an ELO rating of 4,430.
Alpha Zero was able to defeat Stockfish 8, the 2016 winner of the TCEC contest, after about 9 hours of training. Thus, it remains the most powerful chess AI so far, though some claim that the latest version of Stockfish will be able to defeat it.
The major ...
Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.
Read now
Unlock full access