Joint Research Conference

June 24-26, 2014

Machine Learning Web Services In The Cloud - Enabling The Industrial Internet

Abstract:

The industrial internet is expected to stream vast volumes of data into the cloud, taking advantage of the low prices and economics of storage and compute. New tools being built make it easy to develop machine learning models from such data to detect faults, predict failures and anticipate maintenance. The cloud allows such models to be deployed as web services, scaled up to very high transaction rates elastically, and operated with high reliability and availability. This capability will enable a new generation of cloud powered self-diagnosing machines, and manufacturing plants that constantly learn from data about faults and failures. In this talk I will describe the emerging tools
that Microsoft is building to enable such capabilities for machine learning and analytics in the cloud.