Purdue leads Genesis Mission projects in data science and AI

Purdue University faculty will lead Department of Energy Genesis Mission projects aimed at accelerating scientific discovery with AI. Purdue is the only university in Indiana to lead a project under the awards, which selected 278 projects out of 5,000 covering all 50 states across energy, discovery science and national security research.

One Purdue-led effort, directed by Wei Xie, Miaoyuan Liu and Ruqi Zhang, will develop autonomous, on-sensor and real-time machine learning models with agentic AI assistance for the Electron-Ion Collider, a U.S. research facility under construction. The system is intended to identify important particle interactions, compress data and help instruments determine which data matters most while minimizing loss.

A second project led by computer science professor David F. Gleich will study whether large language models can directly cite their training data, a capability described as important for scientific models because it helps researchers understand the basis of AI predictions. Gleich will work with Oak Ridge National Labs, while Xie is partnering with Los Alamos National Laboratory and the Fermi National Accelerator Laboratory.