Machine Learning at the Belle II Experiment: The Full Event Interpretation and Its Validation on Belle Data
Machine Learning at the Belle II Experiment: The Full Event Interpretation and Its Validation on Belle Data
复制标题
Belle II 实验中的机器学习：完整事件解释及其对 Belle 数据的验证
DOI：
10.1007/978-3-319-98249-6
复制
发表时间：
2018
期刊：
影响因子：
--
通讯作者：
T. Keck
中科院分区：
文献类型：
--
作者：
T. Keck
本书探讨了如何利用机器学习来提高昂贵的基础科学实验的效率。第一部分介绍了Belle和Belle II实验，详细介绍了目前许多分析师使用的Belle到Belle II数据转换工具。第二部分涵盖高能物理中的机器学习，详细讨论 Belle II 机器学习基础设施和所选算法。此外，它还研究了几种可用于控制和减少系统不确定性的机器学习技术。第三部分研究了重要的专有 B 标记技术，该技术是在 Y 共振下运行的物理实验所独有的，并深入研究了新颖的完整事件解释算法，该算法将其前身的最大标记侧效率提高了一倍。第四部分介绍了稀有轻子 B 衰变“B→ tau nu”的分支分数的完整测量，用于验证前面部分讨论的算法。
This book explores how machine learning can be used to improve the efficiency of expensive fundamental science experiments. The first part introduces the Belle and Belle II experiments, providing a detailed description of the Belle to Belle II data conversion tool, currently used by many analysts. The second part covers machine learning in high-energy physics, discussing the Belle II machine learning infrastructure and selected algorithms in detail. Furthermore, it examines several machine learning techniques that can be used to control and reduce systematic uncertainties. The third part investigates the important exclusive B tagging technique, unique to physics experiments operating at the Υ resonances, and studies in-depth the novel Full Event Interpretation algorithm, which doubles the maximum tag-side efficiency of its predecessor. The fourth part presents a complete measurement of the branching fraction of the rare leptonic B decay “B→ tau nu”, which is used to validate the algorithms discussed in previous parts.