Modeling charged-particle multiplicity distributions at LHC

Abstract

With many applications in high-energy physics, Deep Learning or Deep Neural Network (DNN) has become noticeable and practical in recent years. In this article, a new technique is presented for modeling the charged particles multiplicity distribution [Formula: see text] of Proton-Proton [Formula: see text] collisions using an efficient DNN model. The charged particles multiplicity n, the total center of mass energy [Formula: see text], and the pseudorapidity [Formula: see text] used as input in DNN model and the desired output is [Formula: see text]. DNN was trained to build a function, which studies the relationship between [Formula: see text]. The DNN model showed a high degree of consistency in matching the data distributions. The DNN model is used to predict with [Formula: see text] not included in the training set. The expected [Formula: see text] had effectively merged the experimental data and the values expected indicate a strong agreement with Large Hadron Collider (LHC) for ATLAS measurement at [Formula: see text], 7 and 8 TeV.