Thesis Defense: Jens Kinch

Title: 

Exploring Machine Learning improvements to the H to Zgamma selection in the ATLAS detector

Abstract:

This thesis outlines the search for using machine learning to improve the selection of the H to (Z to ee)gamma decay in data and simulation from the ATLAS detector. The search covers an initial proof of concept of a multi-stage machine learning architecture, the changes needed in existing software to implement the proof of concept using the existing ATLAS data structures. Following this implementation, within the ATLAS software ecosystem, we will cover a one-to-one comparison to the existing selection. The models are tested on MC and data, yielding an increase in true positives for Z-selection of 26% (19%) in MC (data), and an increase of true positives of 28% for Higgs selection in MC.

Time and place:

Friday 19-09-2025 kl. 14:30-16:30

øv - NBB 2.3.I.164