Maia, the artificial intelligence that plays chess like a person

Must Read

Brian Adam
Professional Blogger, V logger, traveler and explorer of new horizons.
- Advertisement -
- Advertisement -
- Advertisement -
- Advertisement -

Maia is the artificial intelligence engine that plays chess like a person common and not as a supercomputer that tends to beat us easily. This intelligent system is developed by Microsoft in conjunction with researchers from Cornell University and the University of Toronto.

The AI ​​does not try to play chess with the best moves. It learns with each play, victory and defeat, since its learning pattern is similar to ours.

Researchers used millions of online games to train Maia’s intelligence. This allowed the AI ​​to gain experience and behave like a human being when facing its opponents.

Maia is based on the Leela open source system

The project researchers used the Leela open source system to create this artificial intelligence.

Leela is based on DeepMind’s AlphaZero, and is a chess engine that uses algorithms to make the best move in a given position.

Person making a move in chess
Maia relies on the open source Leela system to make the best move in a given position.

This engine allowed Maia to train online and adjust to “different skill levels.” The researchers even developed “nine bots” to engage humans, allowing the AI ​​to augment its abilities.

In fact, the research team claims that Maia’s opponents had scores between 1,100 and 1,900 points, meaning that she faced rookie players.

The AI ​​has faced more than 40 thousand opponents

Artificial intelligence was made available on the chess site lichess.org, where it faced more than 40,000 opponents. Maia started out as a normal rookie, but as she conquered games, her precision and skills increased. In this regard, Ashton Anderson, assistant professor at the University of Toronto added:

“Today’s chess AIs have no idea what mistakes people usually make at a particular skill level. They will tell you all the mistakes you made, all the situations where you couldn’t play with machine precision, but they can’t tell what you should be working on. Maia has algorithmically characterized which errors are typical of which levels, and therefore which errors people should work on and which ones they probably shouldn’t, because they are still too difficult.

Maia, unlike other AIs, does not seek to end us in a game of chess. Its purpose is to learn from us, from our movements, not only to acquire new skills, but also to help us correct our mistakes and improve in the game.

- Advertisement -

Subscribe to our newsletter

To be updated with all the latest news, offers and special announcements.