Analytics and Data Summit 2018

While AI/ML has matured enough to solve many real-world problems effectively, this process typically requires close interactions between three separate skill sets – the domain expert, the data scientist and the programmer. A typical IoT user is a line-of-business person such as a Plant Supervisor or a Fleet Manager. AutoML enables automatic selection of the most appropriate algorithm that suits the domain problem learned based on the interactions with domain expert. Also, AutoML automatically selects the hyper parameters and tunes the models without requiring data scientist skills. Furthermore, AutoML automates the repetitive analytics needed for retraining models warranted by the high velocity IoT data. Therefore, AutoML is the key enabler of analytics in the industrial IoT. Since the typical business benefits in IoT involve significant savings, IoT is the killer use case for AutoML. We will present how Oracle IoT Cloud applies AutoML for industrial IoT problems.