Microsoft has announced the acquisition of deep learning company Maluuba, a Canadian startup that focuses upon a comprehensive view of artificial intelligence, and particularly, the understanding of language. Maluuba’s role in Microsoft is likely to involve fortification of the Cortana digital assistant, as well as other voice-understanding initiatives.
In its blog on the subject Microsoft focused on Maluuba’s integration with Microsoft’s overall AI research efforts. The blog by Harry Shuman, Executive Vice President of Microsoft’s Artificial Intelligence and Research Group Microsoft says:
“Maluuba’s impressive team is addressing some of the fundamental problems in language understanding by modeling some of the innate capabilities of the human brain, from memory and common sense reasoning to curiosity and decision making.”
In addition to Maluuba, the acquisition brings a major deep learning player along from Montréal. Yoshua Bengio is a top expert in deep learning and head of the Montréal Institute of Learning Algorithms. He is an adviser to Maluuba and will continue in an advisory capacity, though his position with Microsoft remains unknown. Notably, Bengio has played a central role in making Montréal a center of deep learning.
Maluuba was founded in 2011 by classmates Sam Pasupalak and Kaheer Suleman from the University of Waterloo, in Canada. The name Maluuba came from Waterloo professor Peter A. Buhr. It was a made up name that he used for fictitious programming languages.
For Maluuba, the acquisition will make it possible to improve AI capabilities generally, and incorporate a range of related technologies. According the co-founders’ announcement:
“So far, our team has focused on the areas of machine reading comprehension, dialogue understanding, and general (human) intelligence capabilities such as memory, common-sense reasoning, and information seeking behavior. Our early research achievements in these domains 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.”
Maluuba also has a revealing insight on its intended directions on its website:
“We finally saw a great opportunity to apply Deep Learning and Reinforcement Learning techniques to solve fundamental problems in language understanding, with the vision of creating a truly literate machine – one that could actually read, comprehend, synthesize, infer and make logical decisions like humans. This meant we had to heavily invest in research, therefore we started our Research lab in Montréal in late 2015 (in addition to our awesome engineering team in Waterloo). Our research lab, located at the epicentre of Deep Learning, is focused on advancing the state-of-the-art in deep learning for human language understanding.”
How this Helps Microsoft
The acquisition of Maluuba continues Microsoft’s activity and interest in the AI sphere since it created the Artificial Intelligence and Research Group last fall.
The acquisition of Maluuba demonstrates the growing AI focus upon understanding and acting upon commands; a capacity demonstrated by IBM’s Watson, and in a simplified but ubiquitous form, by Amazon’s Alexa (we discussed Alexa here). Language is a key to the current development of AI. Language contains not only the basics of communication which might be understood by Natural Language Processing (NLP), but it also contains all of the nuances of human thought. Watson’s victory in the TV Jeopardy competition some years ago increased awareness of this. It is one thing to understand basic language; a secondary problem is to understand well enough to respond; and the third issue is to understand the question and then to be able to formulate a response which meets human criteria of a sufficient answer.
The practical impact of Microsoft’s acquisition is likely to be fairly minimal at present. This is part of Microsoft’s experimental program and will be incorporated in its research toward developing AI capabilities for the future.
Parsing the Pieces
As the AI revolution continues, we are seeing deep learning become ever important in conjunction with other analytics tools to create a more comprehensive AI solution. What we mean by “artificial intelligence” continues to expand as we develop further understanding of the inner workings of both human and machine thought. Some of this is informed by technology, but it also encompasses and interacts with research initiatives in the biological sciences, as well as with the philosophy of intelligence from the past.
It is critical for major software companies to compete in this area. Research is often acquired through mergers with small startup companies such as Maluuba. We can expect further mergers and acquisitions in this area as the technology expands. AI will gradually assert itself in corporate processes; initially through the user interface, and capability to respond to input in natural language. It will move on toward greater comprehension of audio, graphic, and video formats; and then on to the next step of achieving a greater understanding and increasing ability to provide autonomous response.