Search for $t\bar {t}H$, $H \rightarrow b\bar {b}$ at the CMS Experiment in 2016 Using 12.9 fb$^{-1}$ of $pp$ Collision Data

We present an overview of the search for $t \bar t H$ in the CMS experiment using up to 12.9 fb$^{−1}$ of data collected during 2016. The analysis is carried out in the $H \rightarrow b \bar b$ final state with at least one of the top quarks decaying leptonically, resulting in a multi-parton final state with a combinatorial self-background. Discriminators based on machine learning and the direct computation of matrix elements from observed jet and lepton properties are used to distinguish between the $t \bar t H$ signal and the $t \bar t +$ jets background. Using a combined fit of the multivariate discriminants in several event categories, we find an observed (expected) upper limit of $\mu \lt 1.5 \ (1.7)$ at the 95% confidence level. We further discuss how this analysis can be extended to the full Run 2 dataset.