| Description: | Since the Higgs boson was discovered in 2012 at CERN [1,2,3], the research focus is on determining its properties with precision. There are various production and decay modes of the Higgs boson. Using machine learning technology, this analysis addresses the Higgs boson mass reconstruction in the production of the Higgs boson in association with two top quarks, and the Higgs boson decay into a pair of tau-leptons [4,5]. From the decay of the top quarks, there are several particles expected in addition to the decay products of the Higgs boson. In the experiment several objects (electrons, muons, taus, jets) are identified and the challenge is to attribute them to the Higgs boson or top quark decays. There are three analysis levels: event generator level, full ATLAS detector simulation, and real recorded data. Mass reconstruction code was developed and tested on the generator level [6]. In this project, the data from the full ATLAS detector simulation shall be used and the performance of the machine learning algorithms be optimized. The accuracy of the mass reconstruction shall be compared when applied to simulated and recorded data.
Get familiar with the problem and the provided code. Test the existing neural network code, developed on generated truth data, with simulated data of the ATLAS detector. Study and evaluate its performance. |
|---|
