Erlangen 2026 – scientific programme

Parts | Days | Selection | Search | Updates | Downloads | Help

T: Fachverband Teilchenphysik

T 28: Data, AI, Computing, Electronics III

T 28.10: Talk

Tuesday, March 17, 2026, 18:30–18:45, KH 00.024

Machine Learning Models for Separating Signal and Background Events in LHC pp Collisions — Oleksandr Shekhovtsov, André Sopczak, and •Lukas Vicenik — CTU in Prague

We investigate machine-learning-based signal-background discrimination for measuring Higgs boson production in association with top quarks (ttH) in multilepton final states at √s= 14 TeV. We simulated a dataset for a generic detector that mimics a realistic analysis. Low low-level features are used. A range of methods from standard Machine Learning models to advanced approaches inspired by geometric deep learning are benchmarked. The study evaluates these approaches, highlighting their performance and identifying directions for improving symmetry-aware machine learning in collider measurements.

Keywords: Machine Learning; ttH; multilepton