They say good things take time, and that’s exactly the mentality of Toronto’s latest artificial intelligence (AI) research centre, the Vector Institute for Artificial Intelligence.
While the Institute’s official public launch was March 30, employees didn’t take occupancy and start moving in until recently. The last couple months in particular, have been busy. The Vector Institute found a new president and CEO (Garth Gibson) in September and just completed its first round of hiring for research scientists.
“We’ve gone from our original launch event back in March in an empty facility, which was really an announcement by the government that they’re committing to this, to now officially building a functional team, getting it going, finishing our office, and starting to work through all of the things that we said we would do,” Alan Veerman, COO of Vector Institute, tells ITBusiness.ca.
The Institute, located within the MaRS Discovery District, is getting close to $100 million in funding in the next five years from both the provincial and federal governments, as well as approximately $80 million from the private sector over the next 10 years. It is partnered with the University of Toronto (U of T), meaning selected professors and graduate students at the school can further their research using Vector resources as well as outside researchers.
Its original goals were to develop AI for practical, real-life applications, and produce more Masters and PhD degrees in machine and deep learning than anywhere else. After embarking on its journey, the Vector Institute has also made building a strong, sustainable AI ecosystem in Canada its top priority.
“When people ask me what exactly does Vector do, the simplest example I can give is if were a manufacturing plant, then the thing that we would produce is graduates in machine learning and deep learning. But if we don’t do things to help them be relevant to the outside world or to the Canadian ecosystem, then all we’re really doing is putting out smart people with no experience. We’re different from other research institutes in that we are trying to build more linkages into industry,” the COO says.
With the industry facing a severe case of supply and demand problems when it comes to AI talent, the Vector Institute’s goal of producing more experts is likely a welcome goal.
While researchers have the freedom to pursue the areas they want, the Institute has received private sector investment from Canadian e-commerce platform Shopify, as well as all of the top five banks in Canada. Tech giant Google, mass media company Thomson Reuters, global management consulting and professional services firm Accenture, and several investment companies are also all significant funders of the Institute, which gives some idea of where it will focus some of its research.
“We can’t really talk too much about our priorities going forward because we really just learned how to walk after crawling for a few months. We’re now starting to work on how to run, which will involve getting our flagship industry programs underway in 2018 and hiring even more researchers,” Veerman says.
The Vector Institute has two sister organizations in Edmonton, the Alberta Machine Intelligence Institute (AMII), and in Montreal, the Institut québécois d’intelligence artificielle (MILA), that it collaborates with on a frequent basis.
“We’re not competing with [Montreal and Edmonton] at all. This is Canada competing with the US, the UK, Germany, Japan, and other tech countries. I have no problem helping out my colleagues in other cities because at the end of the day, we’re all trying to solve the same problems while racing to out-do other countries,” Veerman laughs.
It hasn’t been all smooth sailing
The biggest hurdle to overcome so far has been convincing world-renowned AI researchers to choose the Vector Institute over other legitimate choices, according to Veerman. He explains that generally, AI experts have three career options: be a university professor/researcher, go work in a big company’s AI lab, or launch a startup.
“The academic route would not pay much but would be fulfilling; a corporate lab would pay significantly better but an expert would be beholden to one corporation, which could limit research opportunities; and startups are full of uncertainty so good luck with that. Maybe you make it, maybe you don’t,” Veerman breaks down. “The Vector Institute is the middle ground where people can research what they feel is important while still working with industry and academia.”
The Institute’s independent governing structure gives researchers the flexibility that most of them desire, he continues.
“If someone wants to be a professor as well as a Vector researcher, that’s great; they can be legally employed by a university and we’ll give them a cross-appointment into Vector to help top them up. If they’re less interested in teaching classes and just want to research, we can become their legal employer and we can give you a cross-appointment into the university to supervise grad students. If you want to work on a startup, you can do that with us,” he says.
It’s also something people could rally around in Toronto because it’s feeding off the AI strength here, Veerman adds. He points out that many of the heads of AI at multinational tech companies like Facebook, Apple, and Google are graduates of U of T’s department of computer science. The city is increasingly becoming a globally-recognized AI research hub, with the likes of Geoffrey Hinton – “the godfather of deep learning” –an Emeritus Distinguished professor at the school, chief scientific advisor at Vector, as well as a lead researcher at Google.
Veerman says the Vector Institute has been on the offensive when it comes to recruiting, using the existing talent in the city to his advantage, as well as arguments like a lower Canadian dollar, high standard of living, and relatively cheaper labour than other high-tech hotbeds.
“We had to openly acknowledge the Canadian context when it came to hiring – that we have talent, but much of that leaves the country. A lot of people think the place to be is Silicon Valley, but we tried to use Canada’s strengths to overcome that,” he highlights. “We also took advantage of the geopolitical instability south of the border – and even across the pond in the UK with all the Brexit talks – and emphasized the government support we’ve received and all the Canadian companies that have large data sets and interesting problems to solve.”
Recent hires include David Fleet, whose research interests include computer vision, image processing, visual perception, and visual neuroscience; Anna Goldenberg, a scientist at Sickkids Research Institute focusing on how to use machine learning for clinical diagnosis of human diseases; and Frank Rudzicz, who is concentrated on natural language and speech processing to improve the quality of life for people with disabilities. There are also several new researchers starting in summer and fall 2018, and you can find the full list on the Vector Institute’s website here.