UBC Undergraduate Research

Exploring Machine Learning Models to Improve the Classification of Displaced Hadronic Jets in the ATLAS Calorimeter de Schaetzen, Rodrigue

Abstract

The Large Hadron Collider (LHC) has yet to find new physics that could address the Standard Model's (SM) large open questions such as the composition of Dark Matter and the matter-antimatter asymmetry of the universe. There have been recent searches for Hidden Sector (HS) particles through the investigation of pair-production of neutral long-lived particles (LLPs) in proton-proton collisions. The ATLAS collaboration recently published results using a partial dataset from a search for paired LLP decays that produce displaced hadronic jets in the ATLAS calorimeter. Several classification models have been studied to identify these LLP decays, including boosted decision trees and LSTMs. In this analysis, 1D convolutional layers were added to an existing model architecture, which significantly improved the performance. Following hyperparameter optimization, the proposed model achieved a ROC AUC score of 0.97; a 10% relative improvement over the previous model.

Item Citations and Data

License

Attribution-NonCommercial 4.0 International

Usage Statistics