The National Science Foundation (NSF) awarded a $199,299 Project Grant under its Computer and Information Science and Engineering (CFDA 47.070) program to the University of California, Berkeley. The grant aims to tailor large language models (LLMs) to enhance personalized learning in computer science education. The research will advance student modeling techniques to trace competencies related to problem-solving strategies and algorithmic thinking. The project will develop LLM-based systems to generate personalized learning scaffolds, such as feedback, worked examples, and suggested problems, using a dual-strategy approach combining LLM fine-tuning and reinforcement learning with human feedback. This research seeks to improve the relevance, competency alignment, and pedagogical helpfulness of the generated learning supports, ultimately benefiting thousands of students across different education levels and broadening participation in computing fields. The award period runs from January 1, 2025, to December 31, 2027.
Generated 3/4/25, 2:33 AM