Rich in AI leaders, Canada tries to finally cash in
So far, our country has not really benefited from technical leadership
TORONTO — Long before Google started working on cars that drive themselves and Amazon was creating home appliances that talk, a handful of researchers in Canada — backed by the Canadian government and universities — were laying the groundwork for today’s boom in artificial intelligence.
But the centre of the commercial gold rush has been a long way away, in Silicon Valley. In recent years, many of Canada’s young AI scientists, lured by lucrative paydays from Google, Facebook, Apple and other companies, have departed. Canada is producing a growing number of AI startups, but they often head to California, where venture capital, business skills and optimism are abundant.
“Canada is not really reaping the benefits from this AI technical leadership and decades of investment by the Canadian government,” said Tiff Macklem, former senior deputy governor of the Bank of Canada, who is dean of the Rotman School of Management at the University of Toronto.
Now bringing AI home is a priority for the Canadian government, companies, universities and technologists. The goal, they say, is to build a business environment around the country’s expertise and to keep the experts its universities create in the country.
And they want to build on the tenacity of veteran researchers like Geoffrey Hinton, Richard Sutton and Yoshua Bengio, who developed techniques that opened the door to remarkable improvements in an AI technology called machine learning, even as many computer scientists and the tech industry considered their work to be an unpromising backwater.
There are encouraging signs, including new government funding, big company investments, programs to nurture startups, and the changing habits of homegrown entrepreneurs and U.S. venture capitalists.
In its recent budget, the government of Prime Minister Justin Trudeau pledged $125 million to support AI research centres in Toronto, Montreal and Edmonton, which will be public-private collaborations.
The Vector Institute for Artificial Intelligence in Toronto, announced two weeks ago, will be one of them. The institute begins with commitments of $130 million, about half the money coming from the national and provincial governments and the other half from corporate sponsors like Google, Accenture and Nvidia, as well as big Canadian companies like the Royal Bank of Canada, Scotiabank and Air Canada.
Hinton, who was hired by Google in 2013 but remains a professor at the University of Toronto, will serve as its chief scientific adviser. The new institute will be in the MaRS Dis-
covery District, a cluster of buildings in downtown Toronto run by a public-private partnership. It is home to many tech startups, including AI companies.
Major technology companies, like Google, Microsoft and IBM, are adding to their AI research teams in Canada.
The experiences of two startups applying AI technology to drug discovery illustrate the challenges — and opportunities — facing Canadian startups.
Atomwise, a company that uses AI technology to predict what new molecules might combat specific diseases like multiple sclerosis, was founded in 2012. Its chief executive, Abraham Heifets, earned his PhD in computer science from the University of Toronto.
When Heifets sought funding, he recalled, one potential Canadian investor said people had tried the same thing 20 years ago. “What could possibly be new?” Heifets said the investor had asked, and turned him down.
Later, Heifets went to the Bay Area and met with Timothy Draper, founder of the venture capital firm Draper Fisher Jurvetson. Draper observed that he had invested in a couple of companies trying a similar approach 20 years ago. That didn’t deter him from trying again.
“That’s a cultural issue, a different appetite for risk and willingness to accept failure,” Heifets said.
Atomwise moved to San Francisco to be close to its investors and the region’s enormous talent pool.
By contrast, Deep Genomics, founded in 2014, has stayed in Canada, and its U.S.based venture backers encouraged it to remain in Toronto.
CEO Brendan Frey studied under Hinton at the University of Toronto, and he has spent years on research that combines deep-learning AI and cell biology. When he hires software engineers, he asks them to make multiyear commitments.
“There are a lot of distractions in the Bay Area,” said Frey, who is also a professor at the university and a co-founder of the new Vector institute. “The hype is a little too hot down there. Besides, we have some of the best talent in the world here.”
Both Atomwise and Deep Genomics were participants in different years in a program called the Creative Destruction Lab. Founded in 2012 by Ajay Agrawal, a professor at the Rotman School, the lab was set up to help technology-intensive startups. They are typically founded by a PhD scientist who has worked on an idea for five years, but has little or no business experience.
In 2015, the program tilted toward AI startups, with 25 companies admitted. Last year, 50 AI startups were admitted, and this year will likely have 75, Agrawal said.
One X factors in Canada’s drive to develop an AI industry is the Trump administration. Canadian AI scientists say they have received a stream of inquiries from U.S. researchers, concerned about the new administration’s stance on immigration and other policies. Should there be a northward migration, it wouldn’t the first time. Hinton settled in Canada in 1987 in part because of the United States’ clandestine support for the Contra guerrillas who sought to overthrow the left-wing Sandinista government in Nicaragua.