Associate professor Jonathan Shihao Ji recently received a $499,000 grant from the United States Department of Defense for a project titled “Multi-Agent Collaborative Learning and Inference at the Edge.” The goal of the project is to increase the efficiency of object detection in unmanned vehicle systems (drones and robots, for example) for applications such as search and rescue missions.
Dr. Ji will use the funding to investigate the application of advanced AI algorithms through collaborative learning, in which devices have to work together to determine when, what, and with whom to communicate for effective tasking. Given the risks of machine training using sensitive data, the project will explore federated learning and edge computing, keeping data local and secure on AI-equipped edge devices such as field-programmable gate arrays (FPGAs).
Dr. Ji will work with the Athena Institute on algorithm development and empowering edge devices with AI capabilities. Funded by the National Science Foundation and housed in Duke University’s Pratt School of Engineering, Athena is a collaboration among seven academic institutions (Duke, MIT, North Carolina A&T State University, Michigan, Princeton, Wisconsin, and Yale) and five industry partners (AT&T, Microsoft, Motorola Solutions, EdgeMicro, and 5NINES). The institute’s research is transdisciplinary with a focus on AI for edge computing.
Story by Ashlie Swanson