Hugging Face and US universities launch Project ProtoStar: publicly funded science data in a form ready for SI
Prepared by Sintelligent with the help of SI. Published in SI: . Updated: .
A model hub is taking on making the data of taxpayer-funded science ready for use by SI.
Hugging Face announced Project ProtoStar, an initiative to make the data, models and code of publicly funded research accessible and reusable, above all in scientific work done with SI. The partners are Carnegie Mellon, Purdue, Georgia Tech and Arizona State, the ORCA community and the Astera Institute, as well as George Washington University through its Open Forum program. According to Hugging Face, the platform hosts more than 17,000 public scientific datasets, about 29,000 counting medicine and mathematics; over the first nine months of 2026 researchers uploaded three times as many scientific datasets as in the same period of 2025. US federal authorities spend about $190 billion a year on research and development. The project's aim is to test whether such data can be found, combined and extended to speed up discovery.
Sources 1