DOE Selects Fermilab for Major AI Research Initiative

Fermilab has been picked by the U.S. Department of Energy (DOE) to fund some projects from the lab in its Genesis Mission. Genesis Mission is a national project by the Department of Energy aimed at exploring the power of artificial intelligence (AI) to aid in advancing scientific research and technology development. The funding goes towards one AI and Machine Learning (AI/ML) project from Fermilab, as well as eight other projects where the lab is acting as a major collaborator.

AI to Improve Particle Accelerator Performance

The key research project conducted at Fermilab is aimed at creating artificial intelligence algorithms that will enhance the work of superconducting particle accelerators. Specialists will design AI-based resonance control systems that are supposed to increase the reliability, efficiency, and affordability of operation of accelerator facilities.

The innovation will enable it to stabilize superconducting radio-frequency (SRF) cavities that play a critical role in accelerators. SRF cavities are very efficient but prone to vibrations and variations in the pressure, which leads to decreased efficiency. AI solutions should help to stabilize the operation of the system, increase the lifetime of radio-frequency amplifiers, reduce the operational costs, and increase the productivity of the facilities. Other organizations involved in the project are national laboratories, universities, and company xLight Inc.

Supporting the DOE's Genesis Mission

Genesis Mission is a DOE program implemented on a national scale and aimed at creating a unified framework, which will integrate the technologies of artificial intelligence, supercomputing, quantum technology, and sophisticated scientific equipment in order to enhance scientific discoveries.

Current grants form the initial stage of the Genesis Mission funding program implementation. Teams of project participants will showcase the ways of using AI-based scientific techniques for better predictions, experiments, and scientific findings.

Fermilab Collaborates on Eight Additional AI Projects

Apart from working on their own AI project, Fermilab will be lending their knowledge to eight additional projects involving artificial intelligence. The eight include analyzing data of Electron-Ion Collider using AI, detection of anomaly in the CMS Level-1 Scouting System, advancing high performance computing, speeding up the Mu2e experiment, supernova identification in real-time for the Deep Underground Neutrino Experiment (DUNE), advancing high-energy physics simulation, advancing neutrino science, and digital twin simulation for fusion magnet systems.

The investment is expected to enable the lab to advance technologies and discoveries in science, according to the Director of Fermilab, Norbert Holtkamp. The CTO of Fermilab, Anna Grassellino, said that the AI control system for the superconducting RF cavities would make the acceleration sites autonomous and efficient, useful not only for particle physics but for many other sciences as well.