Application of Machine Learning to Beam Diagnostics

Machine learning (ML) techniques are widely used in science and industry as a powerful tool for data analysis and automation. Currently, in accelerator physics ML is represented as a young research field, demonstrating mixed results in the latest attempts. The presented work is devoted to exploration of appropriate ML methods for beam diagnostics. The target is to provide an overview of ML techniques which can be applied to improve beam diagnostics and general accelerator performance. Besides the results of ML tools currently used in modern accelerators and evaluation of these tools, we also demonstrate possible concepts with the potential for further investigation and give recommendations on efficient use of ML techniques in accelerators.