Reinforcement learning (RL), a unique learning paradigm inspired by the behaviour of animals and humans, is beginning to show significant promise for controlling and optimising particle accelerators. At Liverpool, Dr. Andrea Santamaria Garcia and collaborators are among the first in the community to explore how these methods can shorten tuning times and open new modes of operation. Building on this momentum, Dr. Andrea Santamaria Garcia has recently led several initiatives that put Liverpool at the forefront of RL for particle accelerators.
Earlier this year, Dr. Andrea Santamaria Garcia was invited to speak at the “International Particle Accelerator Conference” (IPAC) in Taipei, Taiwan, delivering an invited talk on the current challenges of real-world deployment of RL algorithms in particle accelerators. The invited talk was accompanied by the conference paper, “Reinforcement Learning in Particle Accelerators”. At the same meeting, she was also a panellist in the discussion “Future-Proofing Accelerator Operations: Control Systems Meet ML”, offering insights into the opportunities and challenges of deploying AI tools in operational facilities. She was also invited to present her work at the “European AI for Fundamental Physics Conference” (EuCAIFCon) in Cagliari, Italy and the “Particle Accelerators and Beams Conference” in Oxford, UK.
Building on this visibility, Liverpool will also host the next edition of the Reinforcement Learning for Autonomous Accelerators (RL4AA), the fourth instalment of the workshop series. The RL4AA collaboration was started by Dr. Andrea Santamaria Garcia in 2023, with the idea of bringing together the RL and accelerator physics communities to share results, practical insights from the field, and the big questions ahead for real-world RL. RL4AA'26 will take place at the University of Liverpool and the Cockcroft Institute as a LIV.INNO event from the 30th of May to the 1st of April 2026. Expect engaging keynotes, invited and contributed talks, a lively poster session, and a hands-on coding challenge! Further information and registration details can be found on the workshop website.
To support this growing activity, the Cockcroft Institute is advertising an open PhD position on “Goal-Conditioned Reinforcement Learning for Bunch-Shape Control in Linear Accelerators”, which will develop and deploy RL algorithms at the CLARA accelerator in the Daresbury Laboratory. The post is advertised on FindAPhD, with an application deadline of 31st of January 2026. Interested applicants are encouraged to contact Dr. Andrea Santamaria Garcia at ansantam@liverpool.ac.uk for informal enquiries.
In parallel, Dr. Andrea Santamaria is the Liverpool PI for a recently submitted Marie Skłodowska-Curie Doctoral Network (MSCA-DN) proposal, MODERN (Machine-learning Optimised Design of ExpeRimeNts), a European doctoral network on AI-optimised design of large-scale scientific experiments. If funded, this would create a further PhD studentship in RL for particle accelerators as part of a European training network.
Together, these activities mark the arrival of RL as a distinctive new strand of research in the Department of Physics, positioning Liverpool as an active early contributor in this emerging area of intelligent accelerator control.
