Valérie Pisano sips a cappuccino in Montreal’s Caffè Italia, an unassuming gem in the heart of the city’s Little Italy, and a place that she has known since she was a child.
“Coming here is all about history and stability,” said the newly appointed president and chief executive of Mila (formerly the Montreal Institute for Learning Algorithms), the Quebec epicentre of Canada’s artificial-intelligence revolution.
With a soccer match on the TV screen above and the steamy blasts of an espresso machine punctuating her words, Ms. Pisano spoke about the need to stay connected to the past while leading an organization that is literally inventing the future.
“What we’re trying to create is completely new, completely emergent… And as society quickly pivots to this new era, we need to ask, how do we anchor ourselves in the roots of who we are and what we stand for as humanity?”
The future Ms. Pisano sees emerging is just around the corner – not just figuratively, but literally. Mila’s new headquarters is taking shape in a refurbished textile factory a few minutes’ walk away. Situated in the city’s Mile-Ex district, a magnet for hip restaurants and tech entrepreneurs, the building is set to become the world’s most concentrated centre for deep learning and reinforcement learning, two rapidly advancing technical fields that are fueling the AI boom.
Ms. Pisano was brought on in May to serve as ringmaster for the teams of researchers and students who will be shifting from their traditional bases of operation at the University of Montreal and McGill in order to share a space with established companies and small startups seeking to leverage the social and economic potential of AI.
The metaphor is an apt one for the former management consultant, whose CV includes a stint as chief talent officer for Cirque du Soleil.
At Cirque, her role amounted to reinventing the employee experience, helping diverse teams of artists and technicians to come together and create something greater than the sum of its parts. What attracted her to Mila, she said, was the opportunity to bring her interest in human systems to another complex and creative environment – but one that is evolving far more rapidly with consequences for virtually every aspect of society.
“How AI will be integrated into our lives is one of the most significant conversations of our time,” Ms. Pisano said, adding that the chance to be part of that conservation was too enticing to pass up.
A regard for the social impacts of AI is a guiding tenet at Mila, thanks to the institute’s founder and scientific director, University of Montreal professor Yoshua Bengio.
Mr. Bengio is one of the originators of deep learning, a type of programming that allows computers to gain competence through experience and that has already been successfully applied to a broad range of tasks including visual search and speech recognition. Others at Mila specialize in reinforcement learning, a process by which learning systems can explore different possibilities – an automated form of trial and error.
Unlike many colleagues, he has eschewed tying himself to an exclusive corporate relationship and has instead championed the creation of an open research ecosystem around his team. He was the driving force behind a forum held in Montreal last year and its resulting declaration to advance the socially responsible development of AI.
In its new incarnation, Mr. Bengio said that Mila will be testing a vision for how academic researchers and industry can interact in a collaborative way that spurs local economic growth and helps keep top talent in Canada.
“We’re getting into new territory that we’re not used to,” he said of Mila’s new blended environment. A key concern is making sure that the spirit of academic research that has proved so successful in AI up till now is not dampened by close proximity to commercial interests. Over the past two years, tech companies, including global giants such as Microsoft, Google and Facebook, have been busy setting up shop in Montreal in hopes of sweeping up talent developed at Mila. Mr. Bengio has been insistent that the boom unfold on Mila’s terms.
“Because there’s so much demand from companies to work with us we can afford to put the bar pretty high,” he said.
Much rests on getting the relationship right. In June, the Quebec government announced $80-million in financing for Mila over the next five years as a central part of its strategy to develop Montreal’s AI hub. That’s on top of $40-million from the federal government to double Mila’s academic community from 10 to 20 full time professors together with their affiliated students and research associates.
Canada has also placed similar big bets on AI centres in Toronto and Edmonton, but the Montreal cluster stands apart in its spirited determination to be a presence in the community and a voice in the global development of AI.
“They have a real pride in their ecosystem,” said Elissa Strome, a former neuroscientist who oversees the federally funded pan-Canadian strategy on AI administered by the Canadian Institute for Advanced Research, based in Toronto.
Dr. Strome added that among AI centres, Montreal is unique in terms of its rapid growth and diversity, which has fostered startups across a range of application areas from language to heath care, many of them spinoffs from Mila.
Pierre Boivin, CEO of the Bronfman family investment firm Claridge Inc., chairs Mila’s board. While the research organization will continue to run as a not-for-profit in its new setting, he underscored its importance as a catalyst for the AI industry in Montreal. And with the United States, China and Europe all poised to vastly outspend Canada in developing the new technology, Mila’s integrated environment is meant to help Montreal maintain its global standing in AI by operating more nimbly than larger, international centres.
“We believe we’re creating a much faster process for technology transfer,” Mr. Boivin said.
So far, companies are rushing to jump on board, with tenants such as Element AI, Canada’s biggest player in the sector, and Borealis AI, the artificial-intelligence research arm of RBC, preparing to share space with Mila starting in September. Under the current schedule, Mila researcher will start moving into the complex in December.
“If you’re in an ecosystem that’s driven by knowledge, by giving back to the scientific community and leveraging what other people are doing, then co-location makes absolute sense,” said Foteini Agrafioti, who heads up Borealis AI.
Walking around the still empty complex – christened O-Mile-Ex – which consists of two separate structures linked by a bridge, Ms. Pisano pointed out where classes, academic labs and student-led start-ups will be housed in one building and more mature companies in the other. The layout includes a large public space, called the agora, that will encourage interaction.
“The idea is let’s collide – let’s literally bump into each other and get to know each other as human beings,” she said.
The result, one hopes, will defy the science-fiction cliché of artificial intelligence arising from a sterile corporate lab with no regard for human welfare. Perhaps, instead, it will emerge from a quirky and distinctly Montreal mix of art, science, business and culture. Mila sees itself in that second option, Ms. Pisano says, by being more than an accelerator of disruptive technologies. It must also be a thought leader for how that technology can pay back in economic growth to the broader community.
As the world collectively leaps into AI, it will take all the talent and aesthetics of a Cirque du Soleil performance to make sure it lands right side up.