Particle accelerators are some of the most powerful tools that humanity has ever built, supercharging discoveries in physics, chemistry, materials science and biology. The new technologies developed for accelerators have been critical in many modern advancements from producing lifesaving medical isotopes for cancer treatment to breakthroughs in fusion research to eliminating forever chemicals in water.
But particle accelerators are complicated beasts. The most advanced particle accelerators take years to research, design and build, and have tens or hundreds of thousands of devices all working in tandem to deliver a huge variety of particle beams.
“MOAT’s ultimate vision is that we integrate AI so fully into the design, construction and operations of accelerators that we fundamentally transform the pace of discovery and the resulting innovations.”
Jonathan Jarvis, Fermilab
To help address this complexity, Fermilab is playing a central role in the Multi-Office particle Accelerator Team, known as MOAT, which is creating a unified advanced artificial intelligence system that is incorporated into the entire life cycle of particle accelerators to increase efficiency and innovation.
“MOAT’s ultimate vision is that we integrate AI so fully into the design, construction and operations of accelerators that we fundamentally transform the pace of discovery and the resulting innovations,” said Jonathan Jarvis, MOAT collaborator and director of Fermilab’s Accelerator Research Division.
The U.S. Department of Energy’s Genesis Mission is a historic effort to advance AI and accelerate scientific discovery. MOAT is part of the Transformational AI Models Consortium, or ModCon. Fundamental to the mission, ModCon will develop and deploy self-improving AI models that leverage the DOE’s data, facilities and expertise.
Researchers from DOE national laboratories — including Berkeley, Argonne, Fermilab, Jefferson, Oak Ridge, SLAC and Brookhaven — are working together to develop MOAT.
“There are so many applications of accelerators,” said Jean-Luc Vay, head of the Advanced Modeling Program at Lawrence Berkeley National Laboratory and the MOAT project lead. “They have a really big impact across many fields.”
Fermilab’s accelerator technology test facility, called FAST/IOTA, will serve as a key demonstrator for MOAT’s AI tools. FAST/IOTA also offers flexibility for testing across several different types of accelerators and particle beams.
MOAT’s AI systems are still in the early stages of development, but recently MOAT presented the first demonstration of their work to the DOE Office of Science. This showcase highlighted the team’s initial deployment of the Osprey AI tool, which uses AI agents to accelerate specific tasks by a factor of 100. AI agents are autonomous software systems that can reason, plan and take actions with minimal supervision, and they are a key element of MOAT’s approach and long-term vision.
“Usually each of our labs would develop our own standalone prototype,” said Thorsten Hellert, MOAT collaborator at Berkeley Lab and creator of Osprey. “The Genesis Mission has really compelled our community to work together to develop and deploy this new AI software collectively.”
One immediate path to optimization lies in the decades of knowledge created from running a particle accelerator complex like Fermilab’s. Accelerator operators are the heart of these systems, ensuring that the accelerator is running optimally to deliver particles to experiments. When a team responds to an error in the complex, it is critical that they can search for instances of successful problem-solving passed down from previous operators. MOAT’s AI systems will be trained on all of these documented fixes from Fermilab and other DOE accelerator complexes, providing an immediate solution with citations for where the information was found.
“The Genesis Mission has really compelled our community to work together to develop and deploy this new AI software collectively.”
Thorsten Hellert, Berkeley Lab
MOAT will also develop digital twins of each accelerator complex. These will serve as testbeds for virtual diagnostics and speculative beam tuning before any changes are applied. Unlike existing simulations, the virtual twin will be interconnected with the real particle accelerator, allowing for a continuous feedback loop. This enables the AI to learn how the accelerator responds to adjustments and evolve the digital twin to more accurately reflect the performance of the real components inside the accelerator.
MOAT’s AI will be able to be integrated into the conception and R&D phases of accelerators. Fully realized, MOAT’s vision stands to save billions of dollars, years of effort and dramatically increase the performance and value of particle accelerators.
“The goal is for MOAT to speed up how we can discover and expand our knowledge in fundamental physics, chemistry, biology, materials science, and more, faster than would be possible otherwise,” said Vay. “We hope the resulting research would enable us to multiply the research that can be done, whether it’s for new medication, fusion or one of the other particle accelerator applications.”
Fermi National Accelerator Laboratory is America’s national laboratory for particle physics and accelerator research. Fermi Forward Discovery Group manages Fermilab for the U.S. Department of Energy Office of Science. Visit Fermilab’s website at www.fnal.gov and follow us on social media.
MOAT is a unified effort led by the Department of Energy’s Berkeley National Laboratory, in partnership with Fermilab, Argonne, Brookhaven, Jefferson, Oak Ridge, and SLAC national laboratories.
Recommended reading
Early Career Award recipient Aleksandra Ćiprijanović aims to create universal AI analysis framework
Aleksandra Ćiprijanović will use artificial intelligence to solve the domain shift problem in high-energy physics research. Fermilab’s computing capabilities make it uniquely positioned to support this project.
Fermilab drives progress for national AI Genesis Mission
- AI
- artificial intelligence
- astrophysics
- computing
- emerging technologies
- microelectronics
- particle accelerator
- particle physics
- theory
The Genesis Mission is leveraging the strength of the U.S. Department of Energy’s 17 national laboratories, including Fermilab, alongside American research universities and industry partners. The collaborative effort aims to supercharge innovation by integrating the transformative power of artificial intelligence across the national research landscape.
Fermilab’s FAST/IOTA facility achieves major milestone in accelerator research
Fermilab’s FAST/IOTA accelerator test facility has successfully accelerated and stored its first proton beams, marking a major milestone in the lab’s R&D program for next-generation particle accelerator technology. This achievement paves the way for breakthroughs in high-intensity beam physics, artificial intelligence applications and innovations that will power future experiments like the Deep Underground Neutrino Experiment.