Professor Tanja Horn with The Catholic University of America’s nuclear physics group received more than $1 million to fund two research projects integrating artificial intelligence into scientific data collection and analysis.
The U.S. Department of Energy’s GENESIS Mission awarded $300,000 to Horn and co-investigators Dominick Rizk, assistant professor of computer science at Catholic U., and Dmitry Romanov of Jefferson Lab.
The research team is developing an AI-integrated solution that will redefine scientific discovery for Department of Energy facilities, including the Electron-Ion Collider.
According to Horn, the collider will generate continuous streams of detector data approaching 100 gigabits per second. At that speed, massive amounts of data will be produced, making it possible for the new system to identify rare physics signals in real time and prevent data from piling up.
“Existing trigger systems and machine learning approaches are not deployable at this scale due to limited adaptability, poor scalability, and lack of integration with physics constraints,” Horn wrote in the project abstract.
“The AI integration will allow continuously interpreted detector readouts, separate signals from background and prioritize data in real time,” she added, noting that the research is “a key step toward next-generation AI-enabled scientific infrastructure.”
The methodologies will be applicable to other Department of Energy mission areas that rely on high speed, data intensive environments.
The GENESIS Mission is a national initiative that brings together the department’s national laboratories, the National Nuclear Security Administration, academia, industry and other partners. According to the DOE, the initiative aims to develop AI-enabled scientific workflows that could lead to breakthroughs in energy, discovery science and national security.
Horn also received a $750,000 National Science Foundation grant to support research in experimental nuclear physics. The grant will advance the understanding of the fundamental building blocks of matter by incorporating AI and machine learning to accelerate data analysis and improve the efficiency of scientific research.
