Head shot


On October 20, 2023, Brian Plancher, assistant professor of computer science, published new co-authored research in the journal Communications of the ACM, titled “Machine Learning Sensors: A design paradigm for the future of intelligent sensors.” In light of the surge in commercial applications using machine learning over the past decade, and in particular the recent increase in the use of machine learning on embedded devices (TinyML), Plancher and his colleagues outline the MLSensor, a design paradigm to promote a responsible future for intelligent sensors.

The ML sensor is a logical framework for developing ML-enabled embedded systems that empower end users through its privacy-by-design approach. Plancher and his colleagues define an MLSensor as “a self-contained, embedded system that utilizes machine learning to process sensor data on-device — logically decoupling data computation from the main application processor and limiting the data access of the wider system.” By limiting the data interface, the ML sensor paradigm helps protect user data beyond the scope of the sensor’s functionality. And, through the use of detailed datasheets, these sensors can promote transparent and responsible use of ML at the edge. Plancher and his colleagues are optimistic that the ML sensor paradigm will help address some of the challenges presented by pervasive edge machines learning to safely and effectively deploy these technologies into the world.