The Maia project is a human-like neural chess engine. Maia’s goal is to play the human move — not necessarily the best move. As a result, Maia has a more human-like style than previous engines, matching moves played by human players in online games over 50% of the time.
Play against different skill levels against Maia to improve from one level to the next ==> https://lichess.org/@/MAIA
The Maia Project = is a human like neural chess engine
Moderator: Ras
-
- Posts: 5685
- Joined: Wed Sep 05, 2018 2:16 am
- Location: Moving
- Full name: Jorge Picado
-
- Posts: 1945
- Joined: Tue Apr 19, 2016 6:08 am
- Location: U.S.A
- Full name: Andrew Grant
-
- Posts: 1257
- Joined: Wed Mar 08, 2006 8:28 pm
- Location: Florida, USA
Re: The Maia Project = is a human like neural chess engine
It is an interesting project (if a little old). It would be fun to use this type of technology to create an engine that aims to beat a certain rating level in the fewest possible moves. I imagine you could couple Maia technology with a MCTS, so it played lines that were like walking along a precipice for that rating level — where the best move would be difficult for that rating level to find and the most likely moves would lead to disaster. This would be extremely useful for creating opening courses. Currently, the authors of Chessable course will say things like, "this line isn't 100% sound, but it's extremely difficult to play for black if you're under 2000 ELO". The Smith-Mora gambit comes to mind.
Thoughts?
— Steve
Thoughts?
— Steve
http://www.chessprogramming.net - Maverick Chess Engine
-
- Posts: 38
- Joined: Tue Sep 27, 2022 7:13 pm
- Location: United Kingdom, East Midlands
- Full name: Jedidiah F. Sessions
Re: The Maia Project = is a human like neural chess engine
I personally believe that could be quite an interesting experiment.Steve Maughan wrote: ↑Thu Jun 13, 2024 2:49 pm It is an interesting project (if a little old). It would be fun to use this type of technology to create an engine that aims to beat a certain rating level in the fewest possible moves. I imagine you could couple Maia technology with a MCTS, so it played lines that were like walking along a precipice for that rating level — where the best move would be difficult for that rating level to find and the most likely moves would lead to disaster. This would be extremely useful for creating opening courses. Currently, the authors of Chessable course will say things like, "this line isn't 100% sound, but it's extremely difficult to play for black if you're under 2000 ELO". The Smith-Mora gambit comes to mind.
Thoughts?
— Steve