You are here

Reusable Data and Models Will Drive New Advances

WASHINGTON, D.C. - Today, the U.S. Department of Energy (DOE) announced plans to provide $8.5 million for new research to make artificial intelligence (AI) models and data more accessible and reusable to accelerate research in AI research and development (R&D).

The focus is on making data and models reusable by AI application developers and researchers through attention to FAIR Data Principles (Findable, Accessible, Interoperable, and Reusable). The principles—originally proposed in 2016 by an influential collaboration representing universities, industry, funding agencies, and scholarly publishers—were endorsed by world leaders at the 2016 G20 Hangzhou Summit.

The principles are intended to make scientific data more reusable not only by scholars but also by machines, with little or no human intervention. Both capabilities will be important for accelerating the development of AI.

“Data are critical resources for our efforts to use and advance artificial intelligence,” said Barbara Helland, DOE Associate Director of Science for Advanced Scientific Computing Research. “These investments will help us leverage data to advance our use and understanding of AI.”

Proposals are sought in two areas: (1) developing benchmark datasets that incorporate FAIR principles and (2) improving the theoretical understanding of the relationship between AI models and data to promote the development of FAIR models and datasets.

National laboratories and universities will be eligible to lead applications for the two to three-year awards, which will be selected based on peer review.

Total planned funding over three years is $8.5 million in Fiscal Year 2020 dollars.

The DOE Funding Opportunity Announcement, titled “Fair Data and Models for Artificial Intelligence and Machine Learning,” along with a parallel, companion announcement for DOE laboratories, can be found on funding opportunities page of the Office of Advanced Scientific Computing Research within DOE’s Office of Science.