A professor at the University of Alberta is part of a pair presented with what is known as the Nobel Prize in Computing.
Dr. Richard Sutton, a professor of computer science at the University of Alberta, and Andrew Barto, a professor of information and computer sciences at the University of Massachusetts, won the 2024 A.M. Turing award for their work in the research and development of artificial intelligence(AI).
This work includes “developing the conceptual and algorithmic foundations of reinforcement learning,” according to the Association for Computing Machinery (ACM), which presents the award.
“In a series of papers beginning in the 1980s, Barto and Sutton introduced the main ideas, constructed the mathematical foundations, and developed important algorithms for reinforcement learning — one of the most important approaches for creating intelligent systems,” the ACM wrote on the award announcement.
“We were both stunned,” said Sutton. “We were just extremely gratified and surprised and humbled and honoured.”
Their work began when Sutton was a PhD student under Barton and was inspired by observations on psychology, including the “reward” method.
Sutton simplified it as, “if it feels good, do it, and remember it and do it again.”
“We’re coming to understand how we work and how we think,” he said.
The pair published a textbook, Reinforcement Learning: An Introduction, in 1998 that is still considered the standard reference in the field, the ACM added.
“Really, we’re coming to understand what people are, what intelligence is, what minds are, and this is a tremendously important development in the history of intellectual thought.”
Reinforced learning has been used in a number of AI advancements, including ChatGPT and robots learning to solve a physical Rubik’s Cube.
Alberta is set to play a “major role” in the continued development of AI, according to Sutton, in part due to the University of Alberta leading the world in the study of reinforcement learning.
“We’re going to play a pivotal role in the future,” Sutton said.
Ha added that the field still needs to make “fundamental advances,” which could take anywhere from five to 20 years, before the world sees an AI with a mind as good as a human’s.
“It’s always been anticipated that we will have enough computer power to make a human level intelligence by 2030 and that’s still expected. Then the software remains - the ideas, the algorithm, the designs and the agents.”
Specific Reinforced Learning algorithms developed in AI have also provided the “best explanations for a wide range of findings concerning the dopamine system in the human brain,” the ACM said.
Sutton does have concerns about the future of AI, but it’s not something like AI taking jobs, it’s around people’s fear of AI.
“Some things will be economically difficult (in the future), and there’s a danger that we will use AI as a scapegoat we blame that rather than the flaws in our policies, perhaps,” he said.
“All of us have responsibilities for things being used responsibly, AI research is not immune from that, but it is very much a fundamental research area … it’s still early.
“The concerns for application and development, I don’t think it’s appropriate to put those on the fundamental researchers. Those should be on the companies developing them and producing them.”
The award is named after Alan M. Turing, a British mathematician famous for his work in computer science and programming.
The ACM calls the award its “most prestigious technical award.” It also comes with a prize of US$1 million.
