Internet-Based Azure Machine Learning Service Unveiled By Microsoft
June 17, 2014

Internet-Based Azure Machine Learning Service Unveiled By Microsoft

Peter Suciu for - Your Universe Online

On Monday, Microsoft announced it is introducing a new Internet-based service to help companies better analyze data to predict behavior. The Azure Machine Learning service, available for public preview in July, is one of the first of its kind – and is just the latest demonstration of Microsoft's growing focus on machine learning.

Machine learning is part of the artificial intelligence field and it involves constructing systems that can automatically analyze and learn from data. However, the challenge has been how to extract value from the increasing collection of data. Transparency Market Research notes the market for analysis software has been growing about 18 percent a year and will reach $6.5 billion by 2019, Bloomberg Businessweek reported.

Microsoft could make machine learning a little more user friendly.

The Azure Machine Learning Service includes a number of prebuilt model types and packages, which include recommendation engines and decision trees, as well as deep learning models. For businesses, this service can allow users to more easily build algorithms and predict such things as purchasing patterns, tracking future inventory, electricity usage and even the prediction of fraud. The workload and results can be hosted on the web through Microsoft's existing Azure cloud offering.

Azure Machine Learning has been in a trial period for the last year and currently has around 100 customers testing it.

"Azure ML offers a data science experience that is directly accessible to business analysts and domain experts, reducing complexity and broadening participation through better tooling," said early adopter Hans Kristiansen of Capgemini on the Azure Machine Learning website.

Microsoft has said Azure Machine Learning could also dramatically reduce development times to hours or even minutes. It combines Microsoft's own software with open source software that is already publicly available.

"This brings all the benefits of cloud computing to machine learning," Joseph Sirosh, the newly appointed vice president of machine learning at Microsoft, told Cnet's Charles Cooper. "That's a game changer."

Speaking to The New York Times Sirosh added, "This is drag-and-drop software. My high schooler is using this."

Sirosh also said he hopes Azure Machine Learning could open up advanced machine learning to anyone who understands the standard statistical R programming language or even those who have a respectable understanding of statistics.

Microsoft is not alone in developing machine learning technology. It will face competition in this space from cloud computing companies, which include Google and Amazon, as both have developed data frameworks used in building machine learning algorithms and even their own respective analysis services. IBM has also been developing predictive software within its cloud business.

Gigaom reported there are several startups that are developing focused machine learning and artificial intelligence capabilities over API. These include AlchemyAPI, BigML and Expect Labs.

In February, Google bought London-based startup DeepMind to gain an edge in the artificial intelligence market. Clearly a lot of people are now thinking about how machines can learn.

Microsoft has been involved in other artificial intelligence programs, including Skype Translate, a near real-time translation tool, that was officially unveiled last month at the Code Conference.