Project Grant FA23862414011

This $300,000 Project Grant award, funded by the National Science Foundation (NSF) under the Computer and Information Science and Engineering (CFDA 47.070) program, supports research to develop improved evaluation concepts and methods for assessing and enhancing the performance of large language models (LLMs). The key objectives are to: Identify a taxonomy of "evaluation concepts" to pinpoint areas where LLMs are underperforming, such as lack of detail or format issues. Develop...

The University of New South Wales received a $223,350 project grant award from the Department of the Air Force Materiel Command Research Laboratory to support basic research activities under the Air Force Defense Research Sciences Program (CFDA 12.800) from February 1, 2022 through January 31, 2024. The grant funds alternative approaches to investigating stability and transition as proposed in the grantee's September 9, 2021 application titled "BASIC RESEARCH FOR AOARD PROPOSAL 21IOA136 'AN...

This Project Grant award from the National Science Foundation (NSF) under the Computer and Information Science and Engineering (CFDA 47.070) program will develop methods for using large language models (LMs) to apply existing written guidelines, such as grading rubrics or job listings, to decision-making tasks. The $245,862 award will create datasets and annotation tools to study how humans connect guidelines to input tasks, then develop LM prompting and fine-tuning techniques to generate...

The Air Force Research Laboratory awarded a $199,976 Project Grant to the University of New South Wales (UNSW) in Australia under the Air Force Defense Research Sciences Program (CFDA 12.800). The grant, titled "Aligning LLMs for Reasoning and Explanations in Reinforcement Learning," has a performance period from August 1, 2024 to July 31, 2026. The project aims to conduct basic research in support of the AOARD Proposal 24IOA031, focused on developing advanced machine learning...

This $432,656 project grant award from the National Science Foundation's (NSF) Computer and Information Science and Engineering (CISE) program (CFDA 47.070) will support research at the University of Massachusetts (UMass) to explore ways of aligning how artificial intelligence (AI) language models and humans process language. The goal is to better understand the reasons for the discrepancy between how AI and humans handle temporary semantic ambiguity in language comprehension. Specifically,...

This Project Grant award from the National Science Foundation (NSF) Computer and Information Science and Engineering (CISE) program (CFDA 47.070) provides $350,014 to the University of Illinois to develop responsible language models (LMs) with rigorous guarantees. The project aims to enhance the reliability of LMs, which have significantly advanced deep learning but often result in errors in real-world deployment, undermining trust in artificial intelligence (AI) technologies. The research...

This federal Project Grant award from the Department of the Navy's Office of Naval Research (under CFDA program 12.300 - Basic and Applied Scientific Research) provides $125,000.00 to The Trustees of Princeton University for mathematical research on large language models. The award period runs from Sep 1, 2024 to Aug 31, 2027. Princeton University, a prestigious private research university, will conduct advanced scientific research, computational modeling, and interdisciplinary exploration of...

This Project Grant award of $334,731 from the National Science Foundation's (NSF) Computer and Information Science and Engineering (CISE) program (CFDA 47.070) will support research at George Mason University (GMU) to develop language technologies for data-scarce languages. The key objectives are to: Build datasets to explore the meta-linguistic capabilities of large language models. Leverage grammars to build large language models for new languages, similar to how humans learn second languages....

This $875,000 Project Grant was awarded by the National Science Foundation (NSF) under the Computer and Information Science and Engineering (CISE) Federal Grant Program (CFDA 47.070) to the University of California, San Diego (UCSD). The project aims to develop new techniques for aligning large language models (LLMs) with formal specifications in order to generate high-quality computer code that provably matches user intent. Specifically, the project will: (1) develop grammar-aligned decoding...

This Project Grant award from the National Science Foundation (NSF) under the Social, Behavioral, and Economic Sciences (CFDA 47.075) program provides $50,000.00 to Brigham Young University (BYU) to develop biologically-inspired artificial intelligence (AI) language models. The goal is to create AI systems that are more responsive to human emotions, better able to remember and use information provided by humans, and more capable of making fair and equitable decisions when conflicting...