Search for exotic decays of the Higgs boson in 8-jet final state using Neural Network algorithms

This project investigates an exotic Higgs decay channel predicted in BSM scenarios, where the Standard Model Higgs decays into two light scalar a_0 bosons, each decaying into two a_1 bosons, which subsequently decay into pairs of b-quarks—resulting in an 8 b-jet final state. We analyze Monte Carlo simulations to identify discriminating features between signal and background. Using these insights, we train deep neural networks—including CNNs and ParticleNet, following the methodology described in Phys. Rev. D 105, 035008 (2022). Model performance is evaluated on MC to assess its sensitivity to the 8 b-jet signature and its potential to generalize to related signals with 6 or 4 b-jets, then tested on real data to look for possible signal hunt. This work contributes to new strategies for probing exotic Higgs decays.