Project Grant 2534771

The National Science Foundation Division of Computing and Communication Foundations awarded Oregon State University $599,856 on October 1, 2026, under the Computer and Information Science and Engineering program (CFDA 47.070) to develop mathematical foundations and computational methods for identifiable multi-domain data translation and generation using flow matching.

The project addresses two capabilities of modern artificial intelligence: translating information across different data forms (such as converting text descriptions into images) and generating new content with desired characteristics. Recent advances in flow matching—an approach that learns smooth transformation processes between different forms of information—have demonstrated capability in data generation and transformation, but systems cannot reliably guarantee that they preserve intended information during translation or maintain desired characteristics during generation, particularly when trained using unaligned data from multiple domains. The project will establish identifiability guarantees to develop trustworthy and interpretable AI systems less susceptible to hallucinations and unintended distortions in scientific discovery, healthcare, and other data-intensive fields. The work also supports education and workforce development by engaging undergraduate, graduate, and K-12 students in research at the intersection of mathematics, data science, and machine learning.

Performance occurs in Corvallis, Oregon, over a period from October 1, 2026, through September 30, 2029. This is a project grant (assistance type).