At Sensors Expo 2019, embedded processor vendor QuickLogic and their subsidiary SensiML were on a mission to demystify artificial intelligence – without the help of a data scientist.

As opposed to vision or acoustic applications that steal most of the headlines in AI today, SensiML's machine learning algorithms on the 10x other applications that focus almost exclusively on time-series data that you would expect in industrial environments. As shown in a demonstration by SensiML CEO Chris Rogers, test equipment that is typically used to train human engineers can also be used to help train a SensiML classification engine running on an embedded microcontroller.

What you then have is a truly edge AI platform running exclusively at the edge, with none of the latency, security, or network transmission cost concerns of cloud-based machine learning.

Meanwhile, SensiML's parent company, QuickLogic's multicore SoC based on eFPGA technology, an Arm Cortex-M4, a sensor manager, and an audio processing block delivers low-power AI compute for battery-powered applications. Sam Massih, Director of Product Management at QuickLogic, explains how AI algorithms can be run on any of these blocks depending on the use case while other blocks are put into standby mode.

Interested parties can get started on industrial AI without a data scientist today with QuickLogic's QuickAI HDK. Currently, QuickLogic is offering heavy discounts on the QuickAI HDK, which is also compatible with the SensiML stack, for those who order a board at www.quicklogic.com/quickaimercedhdk.