Optimization of ttH¯ Signal and Background Separation Using Machine Learning in the 2lSS1tau Channel

Machine learning techniques have proven the potential to improve the separation of signal and background events in High Energy Physics (HEP) data. This study focuses on the optimization of the ttH¯ production measurement in the analysis channel with two same electrically charged light leptons and one hadronically decaying tau. The analysis is performed with simulated data and the results are expressed as expected sensitivity and expected measurement precision including statistical uncertainties. The performance of binary- and multi-classifier neural networks are compared. In addition, a novel approach combining binary- and multi-classifier neural networks demonstrates further potential to increase the performance