Machine Learning Approaches for BSM Searches: A Novel Strategy for Light Fermiophobic Higgs Detection

by Soo Jin Lee (Konkuk University)

Wednesday 11 Dec 2024, 15:00 → 16:00 Asia/Seoul

CTPU Seminar room

 I will present my research on beyond Standard Model (BSM) physics, highlighting our novel search strategies using machine learning techniques. The main focus will be on our recent work investigating a light fermiophobic Higgs boson (1-10 GeV) in the type-I two-Higgs-doublet model at the HL-LHC. I will demonstrate how we developed an innovative approach using convolutional neural networks to identify diphoton jets, which significantly improved signal sensitivity, particularly for challenging scenarios with heavy charged Higgs bosons. The search methodology we developed can be broadly applied to BSM scenarios where highly collimated photons merge into a single jet, and more generally to studies where both the internal jet substructure and global event characteristics play crucial roles in signal identification.