Search For The Higgs Boson In The Tth H Bb Channel In The Atlas Experiment At The Lhc Using Machine Learning Methods And Synchronization Of The Itk Geometry Description For Simulation And Radiation Studies For The Hl Lhc Atlas Upgrade

The Higgs-top coupling (top Yukawa coupling) measurement can further test the standardmodel, being much stronger than the ones for the other quarks. The associated production of aHiggs boson with a top quark pair (ttH) allows to do a direct measurement. With the ATLASdetector at the LHC, this thesis presents a search for ttH in the H->bb decay mode, rarelyproduced compared with the tt backgrounds. Both processes have final states with multiple jetsand b-jets making the analysis heavily relying on advanced techniques. The large tt modelinguncertainties are a driving factor of the sensitivity.This thesis searches to increase the ttH (H->bb) sensitivity by exploring machine learningmethods. Using early Run 2 data, boosted decision trees are exploited to firstly solve the jet-parton assignment in the reconstruction of the ttH signal, and in a second step classify ttH andtt. The observed significance under the background only hypothesis is 1.4 standard deviations.Targeting a contribution to the analysis round using full Run 2 data, deep learningtechniques are explored: recurrent neural networks as binary classifier solving reconstructionand classification in one step, physics-expertise-aware deep neural networks exploiting rawfeatures, RNN-based multi-classifier for event categorization, and adversarial neural networkaimed to decrease the tt modeling uncertainty.Coping with the new LHC phase starting in mid-2020's, ATLAS will be upgradedto have a new inner tracker. The author contributes to synchronize its geometry descriptionsindependently implemented and used by simulation and radiation studies, which is importantto validate a radiation estimation.