Last year one of the world’s leading research groups in the area of artificial intelligence (AI) announced their first international expansion to Edmonton. The announcement was met with bewilderment by much of the tech news blogosphere, that a company famously bought by Google for $500 million was heading to a northern Canadian city most famous for its connections to the oil and gas industry. For people on the ground, however, the decision made perfect sense. The University of Alberta, after all, is home to one of the world’s most prestigious and important departments for the research of artificial intelligence and machine learning.
The expansion is DeepMind’s first international AI research office, one that will see existing professors working with DeepMind researchers to push the science of AI forward, all while seeking practical applications and connections to local industries and international companies. It has been hailed as a first step in Edmonton becoming a tech centre but the truth is that Edmonton, and Alberta, have been building the perfect home for leading research into AI and machine learning (ML).
“You can point to a critical, distinct moment, which was in 2002, when the province, under the umbrella of Alberta Ingenuity at the time, had a fund to invest in basic science,” Dr. Mike Bowling, one of the university’s professors who will now split his time between his professorship and work with DeepMind, explains. “One of the areas that was funded at the University of Alberta [is now] called the Alberta Machine Intelligence Institute. This was started with provincial funding that they have sustained for almost two decades now. Since then, we’ve been able to almost double in size the number of faculty members doing AI and ML research.”
With sustained, reliable funding from the province, the University has made significant impacts on AI research, famously creating a program that learned how to play heads-up no limit Texas hold’em poker and then beat human professionals. About a dozen graduates from the program ended up at DeepMind in London but now, with Edmonton’s rising AI profile, students and experts are starting to stay and move to Edmonton instead.
AI and Talent Retention
One of the challenges Edmonton has faced with creating renowned AI researchers is convincing those people to stay. “Historically, the students that would graduate would go to Silicon Valley, to London, to New York,” Dr. Bowling explains. “They were going to Microsoft, Google, and Facebook and there was no way they could stay in Edmonton. There weren’t jobs that were going to use the talents that they had. It was hard to stay in Canada and even harder to stay in Edmonton.”
Cynthia van Sundert, executive director of The A100, an Albertan tech entrepreneur and executive group, agrees and says that Edmontonians who left the city are now coming back. “We are starting to see a trend where Edmonton [and] Alberta residents who left home to work have come back to Edmonton,” she says. “These individuals are now building new startups and working on ventures that they intend to keep growing right here. In order to attract, retain and bring back talent, we have to be able to offer interesting and challenging work, and have interesting people working on them. DeepMind will help bring that to Edmonton.”
Van Sundert also says talent retention is just one piece of the puzzle. “Three things are most important right now for Edmonton: talent, capital and profile – not only to help create new ventures but also to help existing ventures scale,” she explains. “Funding is always needed and, thankfully, we have an organization like the Alberta Enterprise Corporation, which was founded specifically to bring venture capital to Alberta [but] Edmonton needs to build its profile as a technology hub. We need to celebrate these successes and let people know that we’ve been here doing all of this for at least a couple of decades now.”
Putting Edmonton on the World Stage
Raising Edmonton’s profile is one of the many jobs for the Alberta Machine Intelligence Institute (Amii). Its executive director, Cameron Schuler, is excited by Edmonton’s potential in AI. “With the resurgence of global interest in AI, we have the potential to build a robust local AI ecosystem and ensure Edmonton is known not only for our excellent research and graduates, but also as a destination for companies seeking to invest in innovation,” he says.
Amii originated with the funding packages made available in 2002 and has since transformed from a purely research-focused initiative to one that connects Alberta’s AI talent with local industry and projects, all while helping researchers secure their funding and forward their research. The strategy, according to Amii’s communications lead Spencer Murray, is threefold: attract and retain talent, find ways to commercialize research, and train companies in AI. “We want to take Edmonton’s AI ecosystem up to the next level,” he says. “We are already seeing successes with people like RBC’s Borealis AI and Mitsubishi Electric, who have established a presence here. We want to increase Alberta’s competitiveness but also to ensure there are diverse receptor sites for the students that are graduating from our schools.”
Connecting Research to Commercialization
Research impact on Edmonton and the world will be easier with DeepMind, Dr. Bowling says, because it is a bridge between the two. “When I think about research impact, I think about the problems that matter 10 years from now. We need people who are thinking about the long-term game. Largely, that has been academic researchers and institutions,” he explains. “One of the nice things about being part of DeepMind is there are always groups of people internally that are looking at these technologies and applying them to individual places. My goals are to be thinking 10 years ahead, and we now have people looking for the more immediate impacts that we can have.”
The impacts are already happening and more are on the way. Google is using DeepMind’s AI to more efficiently cool its servers, saving on energy costs and reducing environmental impact. Similarly, Amii’s recent partnership with ISL Engineering will lead to new efficiencies at the Drayton Valley Water Treatment Plant, where reinforcement learning is helping the plant use its filters to their maximum potential before being cleaned.
DeepMind’s move to Edmonton was ultimately about Alberta’s prestige in the field, but it decided to expand here in part because of the department’s commitment to education as well as research. Dr. Bowling and his fellow professors are passionate about teaching tomorrow’s superstars.
“I love the chances I have to train the next generation of students,” he says. “For all the impact that I can have with pushing the science of AI forward with DeepMind, it would be a shame to lose the other impact I can have, which is to train the next generation of researchers that are going to carry that torch forward.”
DeepMind’s arrival in Edmonton is another step in the city’s growing presence in the world of AI and ML. Their presence will help the many companies and initiatives raise the city’s profile, attract new partners and, just as crucially, help the city retain the talent it is producing. But DeepMind’s arrival is not a story of luck or happenstance, or even the beginning of the story. It is signaling the next act in a story that has been developing for decades.