New Grant Supports AI Innovation in Scientific Archives at the University of Kentucky
An interdisciplinary group of researchers at the University of Kentucky have received a 2026 Celebrating University Research Across the Enterprise (CURATE) award from the Office of the Vice President for Research to advance a groundbreaking project that applies artificial intelligence to natural science archives.
Titled “AI in the Archives: Developing a Machine Learning Tool to Support Natural Science Archival Collections,” the project brings together expertise from rhetoric, computer science, and earth and environmental sciences to develop a machine learning tool capable of identifying connections between archived scientific samples and the research outputs that rely on them. The project includes collaboration with the Kentucky Geological Survey (KGS), specifically its Earth Analysis Research Library (EARL), which houses thousands of geological samples and associated research materials.
The CURATE grant was awarded to Dr. Lauren E. Cagle, associate professor in the Department of Writing, Rhetoric, and Digital Studies, and the project is co-directed by Elizabeth Adams, KGS Director of Engagement, and Dr. Brent Harrison, associate professor in the Department of Computer Science. By analyzing approximately 15,000 pages of publications, theses, and reports known to have used EARL samples, the research team will build a training dataset and keyword library. Funding from the CURATE award will support a graduate research assistant who will manually tag and categorize portions of the dataset, helping to train and refine the machine learning model.
An algorithm will use this dataset and library to identify connections between verified research products and specific EARL samples. Then, a rhetorical analysis of the form and function of those connections will expand the algorithm’s capacity to be applied to other archived scientific samples beyond EARL. The goal is to develop an algorithm that detects when and how archived samples are used in scientific research. This advancement could significantly enhance the accessibility and measurable impact of archival collections worldwide.
“This project has the potential to transform how we understand and document the scientific value of physical collections,” said Elizabeth Adams. “By making connections between archived samples and the research they support more visible and searchable, we’re not only improving access for scientists today, but also ensuring these collections remain relevant and impactful for future discovery.”
In addition to its technical contributions, the project highlights the growing role of digital humanities in scientific research. By incorporating rhetorical analysis into the development of the algorithm, the team aims to create a flexible tool that can be adapted for use by archives beyond EARL, expanding its reach across disciplines and institutions.