Tiny Machine Learning for Energy-Restricted Unmanned Underwater Vehicles
Computer & Information Science Dept.
UMass Dartmouth
Unmanned Underwater Vehicles (UUVs) operate in highly dynamic water environments, constrained by limited power supplies, limited computing capacities, and near-zero communication bandwidth. While data-driven machine learning methods have significantly advanced the field of artificial intelligence (AI) over the past decade, their deployment on UUVs is challenging due to these resource limitations. This talk will initially introduce the concepts of AI agents and environments, and subsequently connect intelligent agents to UUVs. To implement the data-driven deep learning techniques for UUVs, tiny machine learning methods will be reviewed and studied for completing multiple tasks with resource constraints. Our recent research on tiny machine learning will be discussed. Then, the feasibility of developing tiny foundation models for UUVs will be analyzed. The talk will conclude with an overview of future research directions.
Dr. Yuchou Chang is an assistant professor at the Computer and Information Science Department of the University of Massachusetts Dartmouth. His research interests include Artificial Intelligence, Robotics, and Biomedical Imaging. He has authored or co-authored over 100 peer-reviewed publications. His research has been supported by the ONR, NSF, ARO, DOED, and NRC. In 2017, he led a team and obtained the Best Technical Report and Semi-Finalist of the NASA Swarmathon Robotics Virtual Competition. Dr. Chang is a recipient of the E. Kika De La Garza Fellowship at the Department of Agriculture.