Microsoft has unveiled Azure ML, a new machine learning service integrated into the cloud-based Azure platform. This service empowers users to create data-driven applications and APIs for predicting future events. Azure ML leverages machine learning capabilities found in Microsoft’s Bing and Xbox services, utilizing predefined workflows and templates to enhance the speed and usability of machine learning infrastructure.
Joseph Sirosh, Corporate Vice President at Microsoft and the individual overseeing Azure ML, highlighted the potent predictive learning capabilities of machine learning. He emphasized that traditional data analysis enables the prediction of the future, whereas machine learning has the power to alter the future. Sirosh also underlined the role of the cloud in overcoming scalability limitations, describing it as a solution to the “last mile problem” that had hindered progress in machine learning projects.
Sirosh envisions the future of Azure ML as a virtuous cycle between data and APIs. Users bring their data to Azure ML to create APIs, which can be integrated into applications. These applications then generate data that is fed back to the cloud, fueling the creation of more APIs.
While Azure ML is currently in confidential preview, several examples of its applications have been made public. Max 451, for instance, employs it in the retail sector to predict popular products and adjust stock accordingly. Carnegie Mellon University uses Azure ML to predict and mitigate activities in campus buildings, ultimately reducing energy costs.
Azure ML is set to enter public preview in the coming month, with an official release date yet to be announced.
