The George Washington University is a founding partner in Project ProtoStar, a partnership with Hugging Face and other research institutions and universities to accelerate scientific discovery. The initiative aims to make publicly funded datasets, models and code easier for people and AI systems to find, understand and reuse. GW’s participation is led by its Open Source Program Office (OSPO).
Federal agencies including the National Institutes of Health and the National Science Foundation require the research they fund to be publicly available and ask researchers to plan for sharing the underlying data and software. But public is not the same as usable. Much of this material sits in legacy formats, is scattered across incompatible systems, or lacks the documentation that people—and increasingly AI systems—need to build on it.
Through the partnership, GW will set up a presence on Hugging Face for sharing data, code and AI models from GW research. The OSPO will help GW researchers use the platform and will bring ProtoStar’s open science challenges to the GW community. The effort complements work under way at GW Research Computing to support researchers who use open-weight models, including in privacy-sensitive research.
Students will benefit too. Those working in GW research groups will gain hands-on experience with open models and with the open infrastructure at the center of today’s AI development. Instructors will be able to bring the same tools into their courses.
“Public access is a good first step. Usability is next,” said Lorena A. Barba, faculty director of the GW OSPO and professor of mechanical and aerospace engineering. “Open research accelerates discovery when people, and now AI systems, can find it, understand it and build on it. Open source communities have spent decades learning how to make that work for software, and those lessons now apply to data and models. For our students, this is a chance to learn science the way some of the most important advances in AI are happening: in the open.”
Barba, a Jupyter Distinguished Contributor, served on the National Academies committees on Reproducibility and Replicability in Science and on best practices for open code policy in NASA space science. She organized and moderated expert panels on the intersection of open source and AI at the two past GW Open Source Conferences.
GW’s role in ProtoStar grew out of its founding membership in the Open Forum for AI (OFAI), the initiative’s university coalition partner. GW joins fellow founding partners Carnegie Mellon University, Purdue University, the Georgia Institute of Technology, Arizona State University and the Astera Institute. Hugging Face is the infrastructure partner; the Open Research Community Accelerator (ORCA), through its Fostering Innovation Network, is the community partner. Details about the challenges, including how to take part, will be posted soon on the ProtoStar website.
GW researchers and instructors who want to share their work in AI-ready formats, or use Hugging Face in their teaching, can contact the OSPO at [email protected]. The OSPO will host the third GW OSCON on March 31–April 1, 2027.
About the GW Open Source Program Office
Established in 2023 with support from the Alfred P. Sloan Foundation, the GW OSPO is part of Libraries and Academic Innovation. It supports open source software and open science across the university through training, community building and guidance on licensing and sustainable open source practice.
— GW —