Machine learning: adding impact to IoT

Some of the most exciting work being done to reap value from the Internet of Things (IoT) involves taking data insights to the next level using machine learning (ML). A good IoT solution gives you an extraordinarily useful view of your organization’s data; a great one allows you to build algorithms that can predict what’s next—the next equipment breakdown based on historic failure data; the next tune-up needed based on lifespan data from component sensors; the best time to turn on the heat based on last year’s data and next week’s weather report.

It’s a bold new world for IoT—but it’s very real today. Just ask companies like ThyssenKrupp Elevator, which is seeing tremendous new efficiencies by adding Microsoft Azure Machine Learning to its IoT-powered “smart elevators.” Or Pier One, which is putting Azure ML to work on its data to predict what customers will want to buy next, enabling custom-tailored shopping experiences. Or Carnegie Mellon University, which is using Azure ML’s predictive modeling to create a campus full of energy-wise “smart” buildings.

Recently, Microsoft hosted a conference on machine learning, concluding with a panel discussion featuring some of the best and brightest ML scientists around, including Microsoft VP of machine learning, Joseph Sirosh, who underscored the vast potential of ML: “When you put enormous compute against enormous data,” said Sirosh, “and you bring machine learning to bear along with it, and the Internet of Things feeding data into the cloud…and streaming analytics running on live data…I think that in a very short time, you will see a completely different picture of analytics.”  Read a recap of that discussion on the Azure Machine Learning blog.