Neural Network-Based Approaches to High-Energy Physics

Abstract

The exploration of quarkonium states at the Large Hadron Collider (LHC) plays a critical role in advancing particle physics and validating quantum theories. One of the key processes, the decay of J/? mesons into electron-positron pairs (J/??e?e?), offers both valuable insights and challenges, particularly due to the vast datasets produced by high-energy collisions. This study focuses on enhancing the analysis of such collision events through the application of Deep Neural Networks (DNNs).By leveraging techniques such as data preprocessing, feature engineering, and hyperparameter tuning, the study demonstrates the power of DNNs in efficiently processing and classifying complex LHC datasets. The model's performance is assessed using metrics like precision, recall, F1-score, and area under the receiver operating characteristic curve (AUC ROC). The findings underscore the potential of DNNs in improving particle identification and advancing high-energy physics data analysis.