Matt Lakin writes about a collaboration between ORNL and the University of Tennessee, Knoxville that used AI to autonomously steer a synchrotron beamline experiment in real time. The team used the Cornell High Energy Synchrotron Source to map residual strain in a 3D printed metal component, letting the AI decide what to measure next based on incoming data rather than following a fixed plan.
- The beamline was connected to ORNL's INTERSECT infrastructure through the NSF's National Science Data Fabric, enabling remote operation from Tennessee
- The work supports DOE's Genesis Mission, a national initiative to connect computing, data, and experimental facilities into a unified system
- The American Science Cloud, a federated science-optimized environment, is part of the broader infrastructure supporting this effort
- The beamline was connected to ORNL's INTERSECT infrastructure through the NSF's National Science Data Fabric, enabling remote operation from Tennessee
- The work supports DOE's Genesis Mission, a national initiative to connect computing, data, and experimental facilities into a unified system
- The American Science Cloud, a federated science-optimized environment, is part of the broader infrastructure supporting this effort
