Microsoft Aims To Make Machine Learning Mainstream
Many of us look forward to the day when we can get any information we want and have systems intelligently bring us what we're looking for. In a sense that's what Microsoft's new Azure Machine Learning service aims to do. While IBM is among those who have demonstrated the concept with Watson and is looking to advance the technology as well, Microsoft is looking to bring the service to the masses more easily and affordably.
"Simply put, we want to bring big data to the mainstream," wrote Joseph Sirosh, corporate vice president for Machine Learning, in a blog post announcing the general availability for the service that was announced last summer. Azure Machine Learning is based on templates and usual workflows that support APIs and Web services, enabling developers to tap into the Azure Marketplace to easily pull together components to build applications that incorporate predictive analytics capabilities.
"It is a first-of-its-kind, managed cloud service for advanced analytics that makes it dramatically simpler for businesses to predict future trends with data," Sirosh added. "In mere hours, developers and data scientists can build and deploy apps to improve customer experiences, predict and prevent system failures, enhance operational efficiencies, uncover new technical insights or a universe of other benefits. Such advanced analytics normally take weeks or months and require extensive investment in people, hardware and software to manage big data."
Yet despite the rapid growth and rollout of new Hadoop-based services that are the underpinnings of the most sought out predictive analytics platforms, growth is somewhat stalled, according to a survey conducted during Gartner's latest quarterly Hadoop webinar. The percentage of the 1,200 participants who this month said they have deployed Hadoop-based applications has remained flat since last quarter's survey (only 15 percent said they have actually deployed).
However, when the Gartner survey results are examined based on respondents who said they were in the "knowledge gathering" mode, the percentage of Hadoop deployments was lower than 15 percent. Meanwhile, those who said in the survey that they were developing strategies for Hadoop had rates of deployment that were higher than 15 percent. Gartner Research VP Merv Adrian indicated in a blog post that while it's hard to draw any broad conclusions, it may indicate renewed interest by those who have put their plans on hold. "My personal speculation is that it comes from some who have been evaluating for a while," he said.
And indeed there is plenty to look at. Microsoft has rolled out some noteworthy new offerings and is gaining partner support. That includes the latest entry to the Azure Marketplace, Informatica, which released its Cloud Integration Secure Agent on Microsoft Azure and Linux Virtual Machines as well as an Informatica Cloud Connector for Microsoft Azure Storage.
"Users of Azure data services such as Azure HDInsight, Azure Machine Learning and Azure Data Factory can make their data work with access to the broadest set of data sources including on-premises applications, databases, cloud applications and social data," wrote Informatica's Ronen Schwartz, in a blog post. "The new solution enables companies to bring in data from multiple sources for use in Azure data services including Azure HDInsight, Azure Machine Learning, Azure Data Factory and others -- for advanced analytics."
Do you think machine learning is ready for prime time?
Posted by Jeffrey Schwartz on 02/20/2015 at 12:54 PM