AI at the Sensor Level
ABSTRACT
At Neuton.AI, we have developed a revolutionary approach to creating tiny neural networks that consume minimal memory and energy while recognizing complex activities. This breakthrough enables the development of next generation always-on devices with sophisticated functionality and ultra-low power consumption. We are enabling and bringing intelligence to smart sensors at the microwatt level to open up new possibilities for intelligent devices, running for years on a single coin battery.
Neuton.AI's Neural Network Framework creates tiny, accurate ML models without error back propagation or stochastic gradient descent. Our patented algorithm allows the network to grow one neuron at a time, resulting in models under 1 Kb in size. This improves energy efficiency and compatibility with microcontrollers and sensors, while also allowing for more space for valuable business logic. In this session, we'll demo our Human Activity Recognition (HAR) model deployed directly into the sensor. The HAR model recognizes six activity classes with over 98% accuracy. We'll also demonstrate a teeth-brushing tracking solution that recognizes eleven classes, possesses excellent generalization characteristics with an extremely small memory footprint of course. These innovative approaches to building compact neural networks will create more energy-efficient AI solutions for modern IoT products.
BIOGRAPHY
With strong tech expertise and 20+ years of leadership experience, Blair Newman provides unprecedented insights into the future of AI’s development and use. As Neuton’s CTO, Blair is engaged in overseeing our business solutions as well as ensuring high-quality services are delivered to our clients. Prior to Neuton, Blair held various leadership roles at T-Systems North America, a Division of Deutsche Telekom, being responsible for providing strategic direction and leadership in the areas of Dynamic Services (Cloud Computing), SAP Hosting, Application Operations, Managed Hosting and Infrastructure Services.