Learning From The Canadian Model Of AI – Forbes

Learning From The Canadian Model Of AI  Forbes

Canada has received more than its usual share of attention for its AI capabilities. The country was either prescient or lucky in continuing to fund neural networks research when the US retreated from it in the 1970s and 80s. As a result, Canadian researchers like Geoffrey Hinton, Yann LeCun (who is French-American, but worked with Hinton’s group in Toronto), and Yoshua Bengio pushed forward the methods we now call “deep learning.” These three researchers won the 2018 Turing Award—often called the Nobel equivalent for computer science.

Canada is also known in AI for its collegial, public/private ecosystems. They incorporate government funding, venture capital, university research initiatives, and private sector sponsorship. The country has well-developed ecosystems driven by academic research centers now in several cities in Canada. Shelby Austin, who leads Deloitte’s AI practice (called Omnia AI) in Canada, helped me understand how the ecosystems work.

The most prominent research centers are in Toronto (centered around the Vector Institute for Artificial Intelligence), the Montreal Institute for Learning Algorithms (MILA) in that city, and the Alberta Machine Intelligence Institute (AMII) in Edmonton. These three research centers received a $125M CAD grant over five years from the Canadian government for a Pan-Canadian Artificial Intelligence Strategy. Each of these research centers, as well as those in other university-rich Canadian cities like Vancouver and Waterloo, has corporate sponsors. However, the centers and the corporate relationships are relatively new, and like most university research centers around the world, are still learning how best to turn theory into commercial practice.

Canada has already demonstrated that these ecosystems can create ideas for how to develop AI algorithms. Not only deep learning—perhaps the fastest-growing AI technology—but also reinforcement learning and generative adversarial networks (GANs) are algorithm types with strong Canadian roots. The remaining question is whether Canada’s AI structures can develop business applications of AI that lead to commercial success for companies and substantial numbers of jobs in Canada. Granted, big companies like Google, Facebook, Microsoft, and Samsung have opened research labs in Canada and have lured Canadian AI experts to come to work for them. But could there be Canadian firms that make hay from AI?

Austin says that the jury is still out on that issue, but there are several Canada-based organizations that are trying to address it by employing an ecosystem-based approach. In Montreal, the most prominent example is ElementAI, a 2016 startup that is attempting to create a series of AI software solutions. It has close ties with Montreal universities, and Yoshua Bengio, a professor at the University of Montreal, is one of its co-founders. It portrays itself as an alternative to firms like Google and Facebook (employers of Hinton and LeCun, respectively), who are viewed as hiring away academic AI talent.

ElementAI has a set of “fellows”—academics who want to see their ideas implemented in business—who maintain their academic appointments and work a few hours a week for the startup. There are also current students working as interns for ElementAI, and ElementAI employees are involved in a variety of collaborative projects with academic researchers. However, since the company is relatively new and is attempting to build AI infrastructure that other firms can use, it’s perhaps too early to know if it will ultimately be successful.

In Toronto, the Creative Destruction Lab (CDL) at the University of Toronto’s Rotman School of Business is an incubator and accelerator of AI startups rather than a startup itself. Its goal is to accelerate entrepreneurship by providing linkages to mentors and investors for early-stage companies using AI and other emerging technologies. The CDL was established in 2012, and from its beginnings in Toronto has expanded to five other Canadian cities, New York, and Oxford. While it counts 72 AI startups as alumni, it’s still early to say whether it can create successful AI firms.

Large corporations in Canada—its largest banks in particular—are participating in these university-based centers, but several have also created groups of researchers—many of which are also university faculty—within their own walls. Royal Bank of Canada’s Borealis, and TD Bank’s Layer 6, are wholly-owned business units largely comprised of AI researchers who have maintained their university appointments. Some of them are also affiliates of the university research labs we’ve just discussed. Canadian AI ecosystem relationships are rich, complex, and often overlapping.

All of these organizations and their web of relationships are not only a test of Canada’s ability to create large AI firms, but could also provide a model for other countries around the world. The ecosystem approach, in which academics work closely with private sector firms, venture capitalists, and government, is pretty unusual in the United States—at least for AI. It is more common in pharmaceuticals and life sciences, where there are mediating organizations similar to those in Canada that we’ve described.

The ecosystem model is important to success both within and across organizations. To be effective, developing and implementing AI strategy requires the active participation of multiple stakeholders. Business leaders, AI technologists, and policymakers must all play a role. More specifically, organizations looking to implement this model should consider;

·      Organizing: Break down data and digital silos and create platforms that allow many to participate.

·      The right mix: Consider carefully who to partner with externally as part of your ecosystem. Bringing different players together will enhance consumer experience. There is commercial strength in the private sector and large enterprises working together; it’s not about who gets the biggest piece of the pie, the pie is growing exponentially.

·      Turning to talent: Anticipate change by investing in education and research now. Retrain your employees so that they can develop the skills needed to succeed in a digital world. Encourage employees at all levels to work together, and give them the time and space to innovate.

This issue of how to get effective movement of research ideas into commercial practice is an important one for any industry. It’s particularly challenging across the university/business divide. Academics like to work with ideas, and business people like to work with products and services that make money. It isn’t easy to cross that divide successfully and often. If Canada can figure out a way to make it happen with AI, it could be good for that country and the world at large.

Source: forbes.com

Leave a Comment