Artificial intelligence (AI) has the potential to impact all aspects of society: from keeping computer networks secure to detecting nuclear threats. To help the Department of Energy (DOE) Isotope Program understand the potential applications of AI to their research portfolio, Pacific Northwest National Laboratory (PNNL) Chief Data Scientist Draguna Vrabie co-organized the 2022 Workshop on Artificial Intelligence for Isotope R&D and Production and produced a technical report on artificial Intelligence for isotopes.
PNNL is one of several DOE national laboratories tasked with producing isotopes through the DOE Isotope Research and Development and Production Program. Through PNNL’s Isotope Program, Vrabie works to combine AI and machine learning with isotope production.
“Artificial intelligence has the potential to dramatically accelerate breakthroughs in research and development,” said Vrabie. “Integrating AI into isotope science can help us create a safer and more robust production pipeline.”
At PNNL, Vrabie leads the Autonomous Learning and Reasoning team in the Advanced Computing, Mathematics, and Data Division. Vrabie specializes in designing high-performance control systems for cyber-physical systems. Her work in model predictive control for buildings is listed as one of the most cited papers in the journal Annual Reviews in Control.
Vrabie also leads the converged workloads projects in the Data-Model Convergence initiative and the PNNL efforts in the ExaLearn Co-design Center, and is one of the organizers for the upcoming Artificial Intelligence for Robust Engineering & Science (AIRES) 4 workshop at Oak Ridge National Laboratory.
Other PNNL scientists also participated in the Artificial Intelligence for Isotope R&D and Production workshop, including Lab Fellow and Program Development Office Director Karl Mueller, senior research scientist Matt O’Hara, and chief data scientist Court Corley. Mueller and O’Hara gave lightning talks on “Approaching Autonomy: High Throughput Testing, Prediction, and Validation for Redox Electrolytes” and “Improving Separation Method Development Efficiency: Combining Sensor-laden Fluidics with AI/ML,” respectively. Corley gave a talk on “Foundations in AI for Scientific Discovery.”