Garry Kasparov (left) and Murray Campbell (right)
Murray Campbell (’79 BSc, ’81 MSc) travelled a long way from the U of A campus to take his place among the elite team of IBM scientists who designed, engineered, and programmed the first computer to defeat Garry Kasparov—the greatest chess player of all time.
The new chess champion, Deep Blue, was a looming black box designed to calculate 100 million positions per second. The computer boasted impressive processing power— and came about as the result of an immense effort on the part of the IBM team who first lost to Kasparov in 1996.
After their first defeat, Campbell and his team had 15 months to tweak Deep Blue and improve their chances, which meant enhancing the system to compute twice as many positions per second.
Campbell describes the experience of going into the last game of the six-game match. “The last game was tied—there was a lot on the line,” he recalls. Kasparov was playing a risky strategy. “While Deep Blue’s confidence estimates continued to climb, there was always the worry that Deep Blue was misunderstanding something.” Suspended, the creators of the world’s most powerful chess playing computer could only watch as Deep Blue played flawlessly to seal the victory in a brisk 19 minute game.
This iconic moment in history—man vs. machine—captured the world’s attention. Reflecting on the media whirlwind, Campbell notes, “The popular documentary of Kasparov and The Machine was not my favorite film… it was very into the conspiracy theory.” However, he praised the New York City play The Machine by Matthew Charman, which caught the attention of Disney for a future film about the match.
While a triumph like this may seem likely to hold its key players in suspension for all time, Campbell’s scientific momentum continues to take him interesting places. Campbell continues to conduct and contribute to various projects at IBM and has an obvious passion for cognitive intelligence systems and creating smart workplaces.
Much of his current research is more interdisciplinary than some may think when imagining the frontiers of computing science. Some of these projects represent collaborations with scholars from diverse fields; for example, artificial intelligence and machine learning can be used to personalize online educational experiences in MOOCs (massive open online courses). Campbell is quick to point out, though, that each of these innovations still hinge on a strong foundation in artificial intelligence as the basis for success.
Campbell still speaks fondly of his time at the U of A, having completed both a computing science bachelor’s and master’s degree here. In fact, much of his early work in computer chess was conducted under the guidance of the now retired U of A professor Tony Marsland who worked for years on his own software—much of which is still available in the Department of Computing Science.
Campbell expanded on his early research while completing a PhD at Carnegie Mellon University where he created his own chess program, Deep Thought. After completing his PhD, IBM sponsored his research to continue investigating computer chess, eventually leading to the creation of Deep Blue.
With such an influential scientific career behind him, Campbell says that students today should take note that, “there’s been a tremendous return in interest in artificial intelligence, and incredible progress in computers to predict and understand things that weren’t possible 15 years ago. It’s a really good area to get into.”
Good news for U of A students who have the chance to study in the acclaimed program that has contributed influential research on chess, poker, Go, Hex, and other games using research in artificial intelligence. In fact, Campbell places the University of Alberta as “one of the strongest universities in the world for artificial intelligence using gaming as a domain.”
As a member of IBM’s research and development group, Campbell has played a crucial role in research for more than 20 years. In citing IBM’s impact, he notes the creation of Watson, a cognitive artificial intelligence program designed to play Jeopardy. While Campbell is not personally involved in the project, he recognizes it as being on the forefront of cognitive intelligence. “It shows how the field has been developing… it’s interesting to compare the game of chess with the game of Jeopardy. Chess is well defined, Jeopardy is interesting because it brings in natural language, it brings in a question and it can be a difficult question, it can have a joke in it or a pun. These questions can be arbitrarily difficult and then you have to go out and figure out what the answer is by looking at your memories,” explains Campbell. “The distinction is natural language becomes fuzzy and less precise, I think that’s the direction where cognitive science and cognitive computing is moving.”
About the writer
Sarah Beck is a BSc Specialization student in Computing Science; she currently does undergraduate research in interactive narratives and games. In her spare time she is involved with several organizations advocating for diversity in technology and gaming. She contributes to the blog, womeningamestudies.com and tweets from @essefbeck.