Federal Project Grant Award Summary

Northwestern University's Sponsored Research Division received a $420K Project Grant from the National Science Foundation (NSF) Division of Computing and Communication Foundations under the Computer and Information Science and Engineering (CFDA 47.070) program, awarded October 1, 2026, with completion targeted for September 30, 2031. This CAREER award funds research and development activities addressing preference alignment for Large Language Models (LLMs) in open-world environments. The research will produce novel data-efficient preference alignment algorithms designed to enable LLMs to maintain effective alignment with human values when encountering new tasks or domains with limited human-annotated preference data. Additionally, the project will develop algorithmic solutions to enhance the reliability and robustness of LLM-based artificial intelligence systems when confronted with unreliable inputs, including mitigation strategies for data-quality issues such as distribution shifts, label noise, and human value shifts.

The research outcomes are expected to improve data efficiency and reliability of LLMs in high-impact applications including medical diagnosis and molecular discovery. The award integrates research and workforce development activities through educational and training components designed to advance both scientific knowledge in AI safety and human capital development in computing and information science.