This National Science Foundation (NSF) Project Grant award under the Mathematical and Physical Sciences program (CFDA 47.049) supports a project at the University of Utah focused on using advanced natural language processing and large language models to automatically extract materials data from scientific literature. The $200,000 award, made on an EAGER proposal, aims to tackle the challenge of materials data being locked in unstructured PDF formats that are difficult to utilize in modern computational research environments.
The project will leverage the well-curated Pauling File dataset to train the language models to accurately identify and extract key materials synthesis and properties data from research papers. This automation capability could transform solid-state materials chemistry research by enabling wider access to structured data. The project also includes plans for engaging the materials science community and providing educational outreach on the role of AI in materials science. The award period runs from September 1, 2023 to August 31, 2025.
Generated 5/14/24, 1:58 AM