论文标题
喷气标记很容易
Jet tagging made easy
论文作者
论文摘要
我们开发了用于多管齐下的喷气机的标签器，这些喷气机是JET子结构（所谓的“隔离”）变量的简单函数。这些标签者可以以非常简单的方式与喷气质量差异相关。具体而言，我们使用逻辑回归设计（Lord），即使是最简单的机器学习分类器之一，它也显示出超过Atlas和CMS协作使用的简单变量的性能，并且与基于神经网络的更复杂的模型不远。与后者相反，我们的方法可以通过提供已经优化的参数提供简单且可解释的分析公式来轻松实现标记任务。
We develop taggers for multi-pronged jets that are simple functions of jet substructure (so-called `subjettiness') variables. These taggers can be approximately decorrelated from the jet mass in a quite simple way. Specifically, we use a Logistic Regression Design (LoRD) which, even being one of the simplest machine learning classifiers, shows a performance which surpasses that of simple variables used by the ATLAS and CMS Collaborations and is not far from more complex models based on neural networks. Contrary to the latter, our method allows for an easy implementation of tagging tasks by providing a simple and interpretable analytical formula with already optimised parameters.
