Because it needs AI capabilities that are sometimes very specific, Microsoft is not averse to going out and buying technology to meet particular requirements. Such is the case with Maluuba, a startup that is working on applying deep learning to solve problems in natural language processing, according to the Microsoft blog.
Maluuba is a Toronto-based company that has an AI system that is capable of providing text-based reading and comprehension skills that are almost as good as what humans can accomplish. The company has been a competitor of Facebook and Google, which are working on similar systems. Microsoft has a number of applications for Maluuba’s technologies, not the least of which would be making Cortana more human-like at looking at our information — such as email messages — and providing real-world assistance.
According to the Microsoft Blog, Microsoft wants to make its systems as good at reading and writing text as they are at speech and image recognition. “Maluuba’s vision is to advance toward a more general artificial intelligence by creating literate machines that can think, reason and communicate like humans — a vision exactly in line with ours. 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.”
Microsoft presents an example of how Maluuba’s technology could make us more productive, specifically asking an AI assistant based on Maluuba’s machine comprehension capabilities to find the top tax-law experts in an organization. The assistant would be able to scan through the organization’s documents and emails and determine which employee has actually demonstrated the most relevant knowledge.
With Maluuba, we can now expect Cortana to be just as good at reading our information as she is at understanding us when we speak to her. That could make Microsoft’s various productivity tools just as good as us at sifting through our information and finding only the most important bits — as well as doing so proactively and perhaps becoming even better at predicting what we need.
