{"id":759193,"date":"2021-10-19T08:13:55","date_gmt":"2021-10-19T15:13:55","guid":{"rendered":"https:\/\/www.microsoft.com\/en-us\/research\/?post_type=msr-project&p=759193"},"modified":"2023-04-12T14:11:49","modified_gmt":"2023-04-12T21:11:49","slug":"project-maia","status":"publish","type":"msr-project","link":"https:\/\/www.microsoft.com\/en-us\/research\/project\/project-maia\/","title":{"rendered":"Project Maia"},"content":{"rendered":"
\n\t
\n\t\t
\n\t\t\t\"Maia\t\t<\/div>\n\t\t\n\t\t
\n\t\t\t\n\t\t\t
\n\t\t\t\t\n\t\t\t\t
\n\t\t\t\t\t\n\t\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\n\n

Project Maia<\/h1>\n\n\n\n

A human-like neural network chess engine<\/p>\n\n\t\t\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t<\/div>\n\t\t<\/div>\n\t<\/div>\n<\/section>\n\n\n\n\n\n

\"Maia<\/figure>\n\n\n\n

Maia is a human-oriented chess engine that tries to understand human play, rather than optimal play. Whereas existing chess engines ask: \u201cWhat is the best move to play in this position?\u201d, Maia instead asks: \u201cWhat would a human play in this position?\u201d Maia can answer this question for humans of a particular skill level, or even for the specific individual who is playing. By understanding human decisions at a granular level, Maia can help identify a player\u2019s strengths and weaknesses, with the goal of helping them learn and improve their gameplay.<\/p>\n\n\n\n

To learn how Maia works, a good starting point is this blog post<\/a>. For more technical details, please refer to our papers<\/a>. For an overview of Maia\u2019s capabilities and a more hands-on experience, check out this deep dive<\/a>. Because of its fundamentally different approach, Maia feels more human-like than any other chess engine.<\/p>\n\n\n\n

Maia grew out of a simple conversation between MSR researcher Siddhartha Sen<\/a> and Professor Ashton Anderson<\/a> (lead PI of the Maia project, and former postdoc in the Microsoft Research New York City<\/a> lab), who both shared a passion for chess. Given that chess AI has surpassed human abilities since 2005, Ashton and Siddhartha wondered if humans could have a more productive relationship with this AI, rather than simply being beaten by it all the time and being told what move to make without any explanation. In particular, they wondered if this AI could be redirected to help humans, by understanding how they play and showing them how to improve.<\/p>\n\n\n\n

Maia is the first step towards a much larger vision of creating AI that is compatible with humans and synergistic with their goals. We believe that AI can be a partner in advancing human education. To learn more about this vision, listen to this podcast (opens in new tab)<\/span><\/a>. Maia researchers are also active in the chess education (opens in new tab)<\/span><\/a>community an<\/span><\/font>d have been working with chess educators and pioneers to amplify and diversify chess outreach in primary education.<\/p>\n\n\n\n

\n
Play our bots on Lichess<\/a><\/div>\n\n\n\n
Explore Maia Chess<\/a><\/div>\n<\/div>\n\n\n\n\n\n

Why Chess?<\/h2>\n\n\n\n

Chess is an ideal game to study when it comes to artificial intelligence because it is popular, has well-defined rules, and has not yet been fully solved.<\/p>\n\n\n\n

The game emerged in the 15th century and is played between two players controlling the black and white pieces, respectively. Many people know the game and AI researchers use it as a “model system” to study new ideas or techniques.<\/p>\n\n\n\n

\n
\n

What is Project Maia?<\/h2>\n\n\n\n

Maia is a deep learning framework that learns from online human games, with the goal of understanding how humans play. The larger vision of Maia is to use chess to investigate the relationship between humans and AI. Previous AI systems for chess focus on finding the optimal sequence of moves. But it’s more complex to use AI to understand what move a human should make. For example, it’s not always clear that every person will understand a specific move: suggesting an advanced move to a novice player may be dangerous, because the player may not understand the board position that results from that move.<\/p>\n\n\n\n

With this vision in mind, the Maia Project aims to develop an AI engine that holistically understands human play.<\/p>\n<\/div>\n\n\n\n

\n
\n