MSc. Abderahmane Maiza and MSc. Yu Chen

T a l k : 15. May 2025

Machine Learning Based Object Identification at ATLAS

Modern machine learning techniques, has experienced a rapid booming for high energy physics in the last few years. In this seminar, the basic concepts of ML in particle physics are introduced, as well as examples of its application on particle identification, focusing on electrons and jets. A part of the talk will be dedicated to ATLAS flavor-tagging, since it is the most advanced object ID method in terms of ML exploitation. Examples of machine learning models that were applied during Run2 for flavor-tagging will be introduced with focus on explaining their basic concepts, as well as their remarkable role in improving the b-tagging algorithm performance.