Microsoft Banks on Machine Learning to Boost Search, Security
"Mobile-first, cloud-first" may be Microsoft's new mantra, but another term has been has been increasingly creeping into the company's lexicon of late.
As one of the components of Microsoft's growing slate of smart services, machine learning is also guiding part of the company's product strategy, according to Microsoft Research Distinguished Scientist John Platt.
First, it helps to know how his company classifies machine learning (ML). "In general, ML converts data sets into pieces of software, known as 'models,' that can represent the data set and generalize to make predictions on new data," explained Platt in Microsoft's new Machine Learning Blog.
A long-time researcher, Platt is an authority on Machine Learning. He has spent 17 years "creating and using machine-learning algorithms" at the company. And it was time well spent, he suggested.
Platt describes the use of machine learning as "pervasive across Microsoft's vast array of products." He went on to describe the three major ways the tech heavyweight leverages machine learning, namely data mining, statistical engineering and artificial intelligence. The latter discipline uses machine learning "to emulate the human mind, to create computers that can see, hear and understand."
Microsoft credits machine learning with making some its most highly anticipated products possible.
Cortana, which stole the show during this year's Build conference, has roots in machine learning. The company demoed an early version of the digital assistant technology for Windows 8.1 live on stage, putting it through its paces with few hiccups.
Reiterating his comments during the GigaOm Structure Data conference earlier this year, Platt said his company has enlisted the technology to build a better search engine.
"Because Bing has such high volume of traffic, we must use automated decision-making to handle the uncertainty and ambiguity of natural language," stated Platt. "We have exploited ML to create many components of Bing that work together to form a high-quality search engine," he said.
Computer security is another area that his team's innovations are helping improve.
"The Microsoft Malware Protection Center, collaborating with Microsoft Research, has used ML to create software to automatically detect malware, and to help analysts gain insight into malware development," said Platt. The Microsoft Malware Protection Center recently played a role in unraveling the botnets created by the Bladabindi and Jenxcus families of malware that infected millions of PCs worldwide.
Finally, Microsoft is using machine learning to enable computers to "see and hear."
The gesture-recognition software that powers Kinect was created by machine learning said Platt. "Microsoft's speech-recognition system is based on deep learning, a form of ML model that is inspired by the structure of the brain," he added.
Microsoft is working on commercializing the technology in the form of a Skype Translator app that provides real-time text and spoken translation services during Skype video calls. The company plans to release a beta app for Windows 8 sometime this year.