Reinforcement learning is the most difficult learning paradigms to understand and to efficiently use, but it holds a lot of promise in the field of accelerator physics. The applications of reinforcement learning to accelerators today are not very numerous yet, but the interest of the community is growing considerably. This is how the 1st collaboration workshop on Reinforcement Learning for Autonomous Accelerators (RL4AA'23) came to be! The AI4Accelerators team organized and hosted the workshop at KIT, gathering colleagues involved in reinforcement learning. The workshop offered introductory lectures to reinforcement learning, a Python tutorial that studied the real deployment of such an algorithm in a real accelerator, and guided discussion sessions on the most pressing topics. The contents of the discussion will be published in the form of proceedings later.
Most importantly, this the RL4AA’23 event was the official launch of the Reinforcement Learning for Autonomous Accelerators collaboration (RL4AA), which aims to put researchers in contact, create and share valuable resources, and be the spark for interesting ideas and projects together.
Links to the workshop
- Indico page of the workshop.
- Github repository for the hands-on tutorial.
Some event photos
