GNNs for Smart Infrastructure Systems with Large Number of Heterogeneous IoT Sensors
PI Negin Alemazkoor, Ph.D.
This research focuses on efficiently managing the large volumes of IoT sensor data in smart cities and infrastructure systems through tailored machine learning solutions.
While traditional machine learning struggles with this task, Graph Neural Networks (GNNs) offer potential by leveraging data’s inherent graph structure to capture relationships. However, applying GNNs at scale is computationally demanding, prompting the research to develop specialized GNN solutions for infrastructure systems with numerous IoT sensors.
