Project Grant 2523495

This Project Grant award of $598,744.00 from the National Science Foundation's (NSF) Computer and Information Science and Engineering (CISE) program supports research at Rensselaer Polytechnic Institute (RPI) to develop a hybrid AI model that integrates deep learning with probabilistic graphical models. The goal is to create a framework that can effectively incorporate prior knowledge into deep learning models to improve their data efficiency, generalization, and interpretability across applications in manufacturing, scientific discovery, medicine, and defense. The award will systematically categorize and encode diverse forms of prior knowledge, including theoretical and experiential knowledge, using probabilistic graphical models as a unified framework. The project will then develop complementary methods to integrate this encoded knowledge with data within deep learning models at multiple levels. The research will be evaluated on vision-based tasks like human behavior analysis as well as non-vision tasks in scientific discovery and manufacturing. No sub-awards are planned under this grant, which has an award date of September 1, 2025 and an ultimate completion date of August 31, 2028.