机器学习在大型粒子加速器中的应用回顾与展望

机器学习技术在近十几年发展迅猛，并被广泛地用于解决复杂的科学和工程问题。最近十年间，基于机器学习的粒子加速器相关研究也开始呈现出井喷式发展趋势。国际上许多加速器实验室开始尝试用机器学习和大数据技术处理加速器中的海量复杂数据，以期解决加速器及其子系统中的诸多物理和技术问题。不过，迄今为止，机器学习在加速器中的应用仍处于初步探索阶段，不同机器学习算法在解决具体加速器问题的效果及其适用范围尚待摸索，机器学习在实际加速器中的应用仍非常有限。因此，有必要对加速器领域中的机器学习研究做一个整体回顾和总结。将回顾机器学习在大型粒子加速器（以储存环加速器和直线加速器为主）中的加速器技术、束流物理以及加速器整体性能优化等研究方向中已取得的研究成果，并探讨机器学习在加速器领域的未来发展方向和应用前景。

Machine learning applications in large particle accelerator facilities: review and prospects

Rapid growth of machine learning techniques has arisen over last decades, which results in wide applications of machine learning for solving various complex problems in science and engineering. In the last decade, machine learning and big data techniques have been widely applied to the domain of particle accelerators and a growing number of results have been reported. Several particle accelerator laboratories around the world have been starting to explore the potential of machine learning the processing the massive data of accelerators and to tried to solve complex practical problems in accelerators with the aids of machine learning. Nevertheless, current exploration of machine learning application in accelerators is still in a preliminary stage. The effectiveness and limitations of different machine learning algorithms in solving different accelerator problems have not been thoroughly investigated, which limits the further applications of machine learning in actual accelerators. Therefore, it is necessary to review and summarize the developments of machine learning so far in the accelerator field. This paper mainly reviews the successful applications of machine learning in large accelerator facilities, covering the research areas of accelerator technology, beam physics, and accelerator performance optimization, and discusses the future developments and possible applications of machine learning in the accelerator field.