By Katherine Kerr
Martha White would like to dispel the fears and hype about artificial intelligence.
The University of Alberta assistant professor says AI will be a collaborator helping humans make better decisions.
“Once we start to have better and better AI it will be like having a better and better member of our team.… They will have complementary or even similar skills, and together we can all accomplish more,” White says.
White is one of 29 newly minted Canadian Institute for Advanced Research (CIFAR) chairs in Canada and is one of the researchers on the cutting edge of Canada’s drive to dominate in AI research. White and two other Edmonton researchers will receive about $3 million over five years to support their research through the CIFAR chair program. They were nominated by the Alberta Machine Intelligence Institute (Amii). All three are AMII fellows and assistant professors in the U of A’s faculty of science.
White’s focus is on reinforcement learning. Her research aims to understand how AI agents can intelligently interact with their world. The questions she wants to answer, she says, are: “How can we get agents that can make more accurate predictions about the world. And how can we get agents that intelligently explore their world.”
“There’s the long-term goal of understanding intelligence, which maybe isn’t as relevant to helping humans today. But in the short term we are developing algorithms that help us automate decision making. In that way I think we can help humans make better decisions.
“The more information we have, and the better predictions we have about outcomes, the better we can make decisions.”
White says the impression being spread that there are already many advanced AI applications out in the world is overhyped.
“We aren’t very close to having agents that make decisions themselves and tell us how we should behave in the world,” she says.
White says she looks forward to adaptive agents that help humans and that we can have pleasurable interactions with.
“Imagine you could have in-home care systems for elderly people. That would be beautiful.”
“We’re far away because we don’t yet know how to safely or comfortably deploy agents who can make decisions in their world. And a lot of the questions around AI are still fundamental. We haven’t answered the research questions we need to deploy them widely,” says White.
Answering those fundamental questions is the core of White’s work.
The five-year CIFAR grant allows her to be free of chasing yearly grants for her research and to supervise more students. She now supervises 15 students, all of whom have their own projects.
Most of the projects are related to reinforcement learning and machine learning.
White is an unabashed fan of the U of A, Amii and Edmonton as great places for research in AI, machine learning and reinforcement learning.
“The fact that Amii is full of well known people in AI is the reason we have these chairs. There were only three centres that got these chairs: Vector in Toronto, Mila in Montreal and Amii in Edmonton.”
The city is her hometown and where she took her BSc in Mathematics, her BSc in Computing Science, MSc and PhD in computing science. Her PhD, completed in 2014, was supervised by eminent AI scholars Mike Bowling and Dale Schuurmans.
After finishing her PhD, White went to Indiana University Bloomington as a professor before returning to teach at U of A in 2017.
Her husband, Adam White, also an AI researcher, works for think tank DeepMind in Edmonton as well as being a research associate at U of A, teaching courses and supervising students.
And White’s immediate family, including two sisters and a niece are here in town.
White says the U of A is an ideal place for fostering Edmonton’s reputation as a centre for artificial intelligence research.
“I would say the University of Alberta is a very principled place in terms of caring first and foremost about quality in terms of teaching and research and less in terms of numbers. Other universities, particularly in the U.S., care about ‘Make sure you get this much grant money; make sure you get this many papers.’ The U of A is not like that.”
Edmonton also has a thriving ecosystem of the academic and the commercial and some think tank institutions that bridge the pure and applied research gap.
“I don’t want to do commercial things, but I do think impacting the world is important. This is an interesting time when even the professors who are splitting their time with industry labs, — DeepMind, Google Brain, Borealis AI — are continuing to do research in those labs,” says White.
“Amii has an applied machine learning team and that team is encouraged to work with us and do research with us because we want them to continue their training.”
The institute aims to spread its expertise and its students into companies, says White.
White is happy that she can play a particular role as a woman on the faculty and in Amii.
“I think it’s very important we get more diversity into computing science. I would like to attract more women into the field and make them feel comfortable in computer science, so I feel like, as a woman, I get to have a more direct role to do that.”
Amii is set to welcome the next generation of AI experts this summer! Applications are now open for the Deep Learning and Reinforcement Learning Summer School. Hosted in partnership with CIFAR, the Summer School brings together graduate students, post-doctoral fellows and industry professionals to explore the latest AI techniques, build research networks and open collaborative opportunities.
Over the course of 10-days, Edmonton will be home to the brightest minds in a generation and we can’t wait to welcome them to our community.