DOE Genesis Mission awards will advance AI-driven science

Two Cornell-led research teams have been selected to receive nearly $1.2 million through Phase I of the U.S. Department of Energy's (DOE) Genesis Mission, a national initiative that brings together laboratories, universities and industry to accelerate scientific breakthroughs in energy, discovery science and national security using artificial intelligence.

Projects led by Jian-Xun Wang, associate professor in the Sibley School of Mechanical and Aerospace Engineering in Cornell Duffield College of Engineering, and Kyle Shen, the James A. Weeks Professor of Physical Sciences in the College of Arts and Sciences, were selected by the Genesis Mission, which invited interdisciplinary teams to develop new AI models and frameworks to address some of the nation's most challenging problems in areas including energy, advanced manufacturing, biotechnology, critical materials and quantum science.

Wang will lead a $550,000 project, "Differentiable Physics-Integrated Generative Modeling for Complex Turbulent Flows in DOE Energy Systems," with Olivier Desjardins, professor of mechanical and aerospace engineering in Duffield Engineering, along with collaborators at Columbia University and Oak Ridge National Laboratory.

Their project will develop a novel AI framework, known as Physics-Integrated Generative Modeling, to rapidly predict turbulent flows, the chaotic fluid motion that influences the performance, efficiency and safety of technologies ranging from nuclear reactor cooling systems and heat exchangers to advanced manufacturing processes and fusion energy devices.

Rather than running a costly new simulation every time conditions change, the team's AI model will learn once from a library of high-fidelity simulation data, then produce fast, physics-informed predictions while calculating uncertainty and incorporating new observations without requiring retraining.

"This project represents an exciting opportunity to advance a new paradigm for scientific AI by integrating generative modeling with differentiable physics," Wang said. "By developing reusable, physics-consistent foundation models for complex turbulent flows, we hope to accelerate scientific discovery and enable more predictive and trustworthy digital twins for future DOE energy systems."

The second award supports "SPECTRA: Spectral Prediction and Electronic Characterization Through Rapid AI," led by Shen, with co-investigators Eun-Ah Kim, professor of physics (A&S), and Darrell Schlom, the Tisch University Professor in the Department of Materials Science and Engineering in Duffield Engineering. Elio Vescovo of Brookhaven National Laboratory is also a collaborator. The project received $636,000.

The researchers aim to tackle one of condensed matter physics' long-standing challenges: predicting a material's electronic properties before it is ever synthesized.

Researchers often rely on specialized experiments known as angle-resolved photoemission spectroscopy (ARPES) to measure the electronic behavior of materials. Because these experiments are time-consuming, data exists for only a small fraction of known materials.

The team will build the first large-scale, machine-readable database of ARPES measurements and use it to train an AI model that can predict a material's electronic behavior from its atomic structure. The researchers will then compare the AI's predictions with new measurements collected at Brookhaven National Laboratory.

The long-term goal is to accelerate the discovery of materials with valuable technological properties, such as superconductors, by identifying promising candidates before they are ever synthesized.

Phase I awards range from $500,000 to $750,000 and support nine-month projects. Awardees will be eligible to compete for Phase II funding of $6 million to $15 million over three years.

"These awards recognize one of Cornell's greatest strengths: bringing together experts from different disciplines to solve some of society's most complex scientific challenges," said Gary Koretzky, vice provost for research. "By combining advances in artificial intelligence with deep expertise in engineering, physics and materials science, our researchers are developing new approaches that can accelerate discovery while addressing the nation's critical challenges in energy, manufacturing and advanced technologies."

Stephen D'Angelo is the communications manager for biological systems at Cornell Research and Innovation.