Microsoft wants to bring machine learning and the power of predictive analytics to the masses with its new Microsoft Azure Machine Learning service, which it announced today.
To the uninitiated, machine learning may seem like science fiction — after all, it is essentially a branch of artificial intelligence through which systems learn from historical data to predict future behavior or trends. But machine learning is all around us, and its reach is growing.
Machine learning is how search engines learn to provide you targeted results, recommendation engines determine which products might appeal to you, credit card fraud prevention systems seek to protect you from unauthorized charges and map applications help you route around traffic congestion.
Machine Learning for Everyone
But machine learning is also esoteric. It requires expert data scientists and a sophisticated engineering organization. With the new Microsoft Azure Machine Learning service, Microsoft hopes to shatter that reality and build a new one in which machine learning is accessible and affordable for anyone.
“Soon machine learning will help to drastically reduce wait times in emergency rooms, predict disease outbreaks and predict and prevent crime,” says Joseph Sirosh, corporate vice present for machine learning at Microsoft. “To realize that future we need to make machine learning more accessible — to every enterprise and, over time, every one.”
“This tool is intended to democratize machine learning and make it simpler and easier for everyone to use it,” Sirosh adds. “Even my high-schooler son could do it.”
Sirosh notes that anyone comfortable with statistics should find the service accessible, while data scientists will be empowered by replacing a multitude of difficult-to-use tools with a single, easy-to-use tool.
Sirosh says about 100 Microsoft customers are currently testing the Microsoft Azure Machine Learning service, and the company will open the service to public preview in July. The service is a fully managed cloud service for building predictive analytics solutions. It will feature visual workflows and startup templates for common machine learning tasks, many of them based on the algorithms Microsoft has developed for its own products like Xbox and Bing.
Machine Learning Service Will Use App Store-like Model
You will also be able to publish APIs and Web services through the Azure-based offering.
“There’s going to be a wonderful ecosystem that’s going to develop around this,” Sirosh says. “People are going to find those APIs incredibly useful. Being in one central place, hosted in the cloud, users will be able to share experiments, share data, and the systems will constantly improve with usage. As more and more people use it, the service itself will become more and more rich.”
Sirosh says the service will also lead to a marketplace in the cloud, as data scientists and statisticians with expertise in specific areas will be able to create APIs and charge for them.
“That’s an incredible app store-like model that allows people to create applications and monetize them in the cloud,” Sirosh says.
Additionally, Sirosh notes that the service will also empower Microsoft partners, systems integrators and consultants, making it easier for them to help their customers stand up predictive analytics.
“There will be online training associated with this on the cloud itself,” he says. “We have training partners, we have a partner boot camp. For reach to the broad audience, you will see more and more courses and free material put up in the cloud. Hopefully, this will become the preferred place to learn about data science.”
Microsoft has not yet shared pricing for the Azure Machine Learning service, but Sirosh says it will follow a usage-based model indexed to compute and the number of transactions that go through APIs. There will also be a free tier.
Follow Thor on Google+
Thor Olavsrud covers IT Security, Big Data, Open Source, Microsoft Tools and Servers for CIO.com. Follow Thor on Twitter @ThorOlavsrud. Follow everything from CIO.com on Twitter @CIOonline, Facebook, Google + and LinkedIn.