Microsoft Corp. is buying into Canada's artificial intelligence research community with a deal to purchase Maluuba, a deep-learning and AI– technology startup based in Waterloo, Ont., and Montreal.
The company, founded in 2011 by two University of Waterloo grads, focuses on teaching machines to read like a human brain – a massive computer science challenge that has bedevilled researchers for decades. Early on, the company also developed language-recognition software for voice assistants in smartphones and cars.
The company raised $9-million in venture funding in January, 2016, but neither company released details Friday about the sale price. A statement from a Microsoft spokesperson didn't rule out possible layoffs among the 50 current employees. "Microsoft representatives met with Maluuba employees earlier today in both Waterloo and Montreal. We are currently working on integration plans and expect most, if not all, employees will join Microsoft."
In the statement announcing the acquisition, Maluuba co-founders Sam Pasupalak and Kaheer Suleman said the rationale for selling was to bolster resources.
"Early research achievements [in machine-reading comprehension, memory, common-sense reasoning] accelerated our need to scale our team rapidly; it was apparent that we needed to bolster our work with significant resources to advance towards solving artificial general intelligence."
Microsoft also announced Friday that Yoshua Bengio – one of the leading lights on deep learning as well as the head of the University of Montreal's Institute for Learning Algorithms – would become an adviser to the company's Artificial Intelligence and Research Group, much as he had been advising Maluuba.
Maluuba makes much of its research available to other AI developers. In November and December, it released massive datasets that assist with question-and-answer scenarios as well as a product that helps machines remember and contextualize a string of dialogue. Just a month ago it partnered with McGill University's Reasoning and Learning Lab, a collaboration that received an NSERC Engage grant from the federal government.
Canadian machine-learning researchers have been hot properties as the world's biggest companies seek to make devices and software more intelligent. In 2013, Google hired deep learning and neural network pioneer Geoffrey Hinton from the University of Toronto. Former University of Toronto machine-learning professor Ruslan Salakhutdinov (more recently of Carnegie Mellon University) was hired in October by Apple to help it improve its Siri voice assistant.