Uncovering hidden new physics patterns in collider events using Bayesian probabilistic models
Uncovering hidden new physics patterns in collider events using Bayesian probabilistic models
复制标题
使用贝叶斯概率模型揭示对撞事件中隐藏的新物理模式
DOI：
--
复制
发表时间：
2021
期刊：
影响因子：
--
通讯作者：
D. Faroughy
中科院分区：
文献类型：
--
作者：
D. Faroughy
在像LHC这样的高能对撞机上发生的单个事件可以用一系列测量或“点模式”来表示。从这个通用的数据表示，我们建立了一个简单的贝叶斯概率模型的事件测量有用的无监督事件分类超越标准模型（BSM）的研究。为了达到这个模型，我们假设事件测量是可交换的（并应用De Finetti的表示定理），数据是离散的，测量是从多个“潜在”分布（称为主题）生成的。由此产生的碰撞事件的概率模型是一个混合成员模型，称为潜在狄利克雷分配（LDA），广泛用于自然语言处理应用程序的模型。通过对QCD和BSM的混合双喷流样本进行训练，我们证明了双主题LDA模型可以学习区分（未标记的）喷流子结构数据中隐藏的新物理模式，这些模式由非平凡的BSM签名产生，来自更大的QCD背景。
Individual events at high-energy colliders like the LHC can be represented by a sequence of measurements, or ‘point patterns’. Starting from this generic data representation, we build a simple Bayesian probabilistic model for event measurements useful for unsupervised event classification in beyond the standard model (BSM) studies. In order to arrive to this model we assume that the event measurements are exchangeable (and apply De Finetti’s representation theorem), the data is discrete, and measurements are generated frommultiple ‘latent’ distributions (called themes). The resulting probabilistic model for collider events is a mixed-membership model known as Latent Dirichlet Allocation (LDA), a model extensively used in natural language processing applications. By training on mixed dijet samples of QCD and BSM, we demonstrate that a two-theme LDA model can learn to distinguish in (unlabelled) jet substructure data the hidden new physics patterns produced by a non-trivial BSM signature from a much larger QCD background.