• Learning in Machines & Brains

AI success holds innovation lessons for Canada

by Alan Bernstein
Apr 23 / 18

Last week CIFAR’s board of directors met in Montreal.

Alan BernsteinThat evening, following the board meeting in Power Corporations beautiful boardroom, we held a dinner for close to 100 people from academia, government, and the private sector. Following the dinner, there was a panel discussion around the theme CIFAR & Quebec: An AI Success Story. Graciously hosted by CIFAR board member Olivier Desmerais, the event celebrated the vibrancy of Quebec’s ecosystem in artificial intelligence (AI), and the great partnership between CIFAR and Quebec in establishing that ecosystem.

The panel featured Rémi Quirion, Chief Scientist of Quebec; CIFAR Senior Fellow and Program Co-Director Yoshua Bengio of MILA and the Université de Montreal; CIFAR Senior Fellow Joelle Pineau of McGill University and Facebook AI Research Lab; Pierre Boivin, President & CEO of Claridge Inc. and MILA Board Chair; and Philippe Beaudoin, Co-Founder of Element AI.

Montreal’s and Quebec’s excellence in AI is obvious. From MILA, to the centres of excellence at the Université de Montréal and McGill to a highly robust ecosystem that includes giants in the field like Facebook and Microsoft, scale ups like Element AI and start ups that are too numerous to mention. Without any question, Montreal has established itself as a global center of excellence for AI not just in Canada but the world.

This global excellence didn’t just happen. It is the result of many years of people and organizations working together to realize a shared vision and common goals. In my remarks that night, I noted that we should not take this success for granted. Indeed, Montreal’s (and Canada’s) success in AI research and innovation stands out as an exception in what is otherwise Canada’s ongoing poor performance, as reported in this year’s Council of Canadian Academy’s Report, to turn research into new products or successful companies.

I believe Canada has much to learn from our exceptional success in AI research and innovation. First, CIFAR’s support of AI research, going back to the 1980s during what was then called the AI winter, has been key to attracting, retaining and training AI talent. Led initially by Geoff Hinton, CIFAR Distinguished Fellow from the University of Toronto, and now by Yoshua Bengio and Yann LeCun of New York University, that program has undoubtedly the best collection of AI talent on the planet.

Element AI’s Philippe Beaudoin, talking about how he saw CIFAR, said it best when he said:

“What CIFAR got right was the fact that by connecting great academic minds together, this is where the good ideas come from. It’s really easy to think its only about the funding. But at some point, the extra million dollars in funding is worth less than giving time to brilliant people who are willing to collaborate, to sit in the same room and come up with and solve great research questions.”

Great scientists have to be supported by great institutions. The universities in Montreal, Toronto and New York and others that have supported Geoff, Yoshua and Yann and their colleagues deserve great credit for the support they have provided and continue to provide.

“What CIFAR got right was the fact that by connecting great academic minds together, this is where the good ideas come from.”

Governments are also key in providing long-term and internationally competitive levels of support for fundamental research. Governments are the true high-risk angel investors in fundamental research. CIFAR is grateful that, almost since our very beginnings over 35 years ago, the Canadian government and many provincial governments, including the governments of Ontario and Quebec, have generously supported us. That support has allowed CIFAR to do what we set out to do from our very beginnings; viz., to address some of the most important challenges of our time by bringing together the world’s best researchers, regardless of discipline or country. Public sector support, matched several times over by generous philanthropic support, has been essential to our success.

Risk taking, great institutions, and exceptional scientists are all key. But independently, that is not enough. As our panelists emphasized, the key to Canada’s and Montreal’s success in AI has been alignment, the willingness of all the moving parts in this ecosystem to align for common purpose. That sense of community has created the synergy and excitement that has attracted more talent, has attracted firms like Google, Google Deep Mind, Facebook and Microsoft, pools of capital like Real Ventures and Caisse de dépôt, and encouraged Canadian companies like RBC, TD and Magna to invest. And the creation of new organizations like MILA serve as a critical meeting place for all these moving parts. And successful scale-up companies like Element AI have sent a signal to students, government, and business that Quebec and AI is an attractive place to invest your time, talent and money.

The pan Canadian AI Strategy has three additional goals: create Canada CIFAR AI chairs to attract, retain and train top AI scientific talent, develop a national AI program and join a global conversation on the ethical, legal, social and economic issues around the applications of AI.

For 35 plus years, CIFAR has been a convener, not just of great researchers but also of great organizations., And so when the Canadian government sought to develop a pan-Canadian initiative that would strengthen Canada’s position as a leader in AI research and innovation, they naturally turned to CIFAR to develop and implement what has become the $125 million Pan-Canadian AI Strategy. We are now working closely with MILA in Montreal and their sister organizations, the Vector Institute in Toronto and Amii in Edmonton. Just as with MILA in Montreal, these institutes are hubs of AI expertise that are serving as centres of gravity for the AI ecosystems in those three cities.

In addition to providing financial support for the three institutes, the pan Canadian AI Strategy has three additional goals: create Canada CIFAR AI chairs to attract, retain and train top AI scientific talent, develop a national AI program and join a global conversation on the ethical, legal, social and economic issues around the applications of AI.

The creation of these three institutes is an important experiment that goes far beyond AI. CIFAR’s hope is that, like the two-faced Roman god Janus that served as a doorway straddling two rooms, the three institutes will serve as an important experiment in how to turn great ideas coming from our talented academic scientists and scholars into innovation in the public and private sectors.

Too often in this country we bemoan our poor innovation performance in the global race to turn research into new companies, new products or processes. There are three important lessons from Quebec and Canada’s success in AI: First, great innovation starts with great science. Second, only the public sector can support the high-risk, long-term fundamental research that is the fodder for innovation. And only the private sector can turn great new ideas into new companies and products. Third, it is this complementarity and alignment between all the players in the research and innovation ecosystem, built on a foundation of true international excellence in science, that will be key to our success in innovation.