Designer Track Invited

Session 43: Machine Learning at the Edge

10:30 AM–12:00 PM Jun 27, 2018 (America - Los Angeles)

Room 2010

As machine learning spreads across all cloud service layers, IoT devices at the edge are the next battle ground for intelligent applications that leverage local sensor data in real time, with high availability and reliable privacy.  In this session we will cover the challenges of collecting sensor data from multiple sources, scaling the number of edge nodes, delivering low latency data communications, extracting data features and fusing them. We will present modern machine learning techniques to process the acquired data with a particular focus on scaling Deep Neural Networks from very low power microcontrollers to application processors and the power of knowledge transfer between nodes.