Federal Project Grant Award Summary

The National Science Foundation's Division of Computing and Communication Foundations (CFDA 47.070) awarded $282,156 to the University of South Carolina on September 1, 2026, for a five-year CAREER grant to develop novel artificial intelligence (AI) algorithms that solve pathfinding problems based on high-level goal specifications. The project, which extends through August 31, 2031, addresses a critical challenge in domains ranging from robotics to chemical synthesis: enabling AI systems to reach goals defined by properties rather than explicit state descriptions. The research will produce domain-independent algorithms that integrate machine learning, heuristic search, and formal logic, utilizing deep reinforcement learning to train neural network heuristic functions and developing neuro-symbolic refinement methods that learn from failure cases to improve performance.

Beyond algorithm development, the award supports educational and training outcomes through provision of hands-on experience for students in AI research. The deliverables from this CAREER award—which emphasizes research excellence and early-career faculty development—will advance the state of practice in AI-driven pathfinding applications while building workforce capacity in computing and information science. The research outcomes are expected to reduce time and cost requirements for solving complex pathfinding problems and enable discovery of novel goal configurations across multiple scientific and industrial domains.