University of Alberta (AI research)

What it is

The University of Alberta is a university whose Department of Computing Science has made game playing a central part of its artificial-intelligence work. The department hosts the official project page for Chinook, the checkers program built by a university team, on its own site rather than on a personal homepage ^[1][3]. Chinook was built at the university by a team led by Jonathan Schaeffer, who was then dean of science ^[3].

Putting the project page on a departmental site is a small signal of how the work is treated: as an institutional result that the department keeps available, not as one researcher's private side project ^[1].

Why games

Richard Sutton, a computer scientist at the university, describes games as a simplified system for testing theories, the way biologists use fruit flies ^[3]. The point is that a game has explicit rules, a definite outcome and none of the noise of the real world, so an idea about how a machine should search through possibilities or learn from experience can be tried out cheaply and judged without ambiguity. A worked example makes the idea concrete: a program that plays checkers has to choose among legal moves, evaluate positions it cannot fully resolve, and improve as it plays. Whatever is learned about doing that well is a claim about machine reasoning in general, not only about checkers.

The limitation is obvious and worth stating: a game is still a closed world. A method that wins on a board does not automatically carry over to open-ended problems such as language, perception or planning in a physical environment.

The checkers work

Schaeffer says copies of his checkers program started running in 1989 on undergraduate machines at the University of Alberta ^[2]. That detail is part of the story the university tells about the work: it began on ordinary student hardware, at a time when the machines available to a department were far less powerful than the ones used for headline results later on.

The people

Schaeffer was retiring in 2024 after a 40-year career as a professor in the Department of Computing Science ^[2]. The university's publication The Quad describes him as a pioneer of artificial intelligence at the University of Alberta and a fellow of the American Association for Artificial Intelligence and of the Royal Society of Canada ^[2].

His administrative work shaped the department's surroundings as much as his research did. As a past dean of the Faculty of Science, he expanded the university's reputation as a leader in AI and led its development of massive open online courses ^[2]. He also co-founded the Alberta Machine Intelligence Institute ^[2], and he supervised over 80 graduate students ^[2] — a large group of people trained in the same department and, in many cases, on the same kinds of problems.

The university's connection to game-playing AI reaches beyond its own campus. Murray Campbell, who co-led the IBM Deep Blue team, is also part of the U of A connection ^[3].

Where it sits

Taken together, these pieces describe a department that built its reputation around programs that play games and around the people who train them. The Chinook page is the visible artefact ^[1]; the institute Schaeffer co-founded, the online courses he led, and the students he supervised are the institutional layer underneath ^[2]; and the Deep Blue link shows how far that line of work travelled ^[3]. Games, in Sutton's framing, are the fruit flies — small, controllable organisms that let researchers ask large questions ^[3].