Speaker
Description
At the Accelerator Test Facility (ATF) at KEK, research and development of beam control and diagnostic techniques is being carried out toward the realization of nanometer-scale beams required for the International Linear Collider. Since nanobeam tuning requires the adjustment of many parameters under limited machine time and human resources, efficient optimization methods are essential. In this work, we have developed and implemented a beam tuning method based on Bayesian optimization. The results confirm more efficient tuning than conventional manual operation. In parallel, a big-data analysis framework is also being developed to organize and utilize operational data for improving machine understanding and future automated tuning. This presentation gives an overview of these machine-learning-based developments at ATF and discusses their potential for efficient beam operation.
| I have read and accept the Privacy Policy Statement | Yes |
|---|