Study of $\mathrm{B_s^0}$ meson reconstruction in proton-proton collisions using machine learning with ALICE

Measurements of the production of heavy-flavor hadrons containing charm or beauty quarks in proton-proton collisions provide an important test for pertubative quantum chromodynamics and allow to study heavy-flavor hadronisation mechanisms. In Run 2, the ALICE collaboration studied the production of beauty hadrons only indirectly via D mesons originating from beauty-hadron decays. By using software triggers for the selection of events containing heavy-flavor hadrons measured by ALICE in Run 3, direct measurements of the production of beauty hadrons and the $\mathrm{B_s^0}$ meson in particular are expected to be feasible. In the course of this summer student project, the complete workflow for the reconstruction and selection of $\mathrm{B_s^0}$ candidates in $\mathrm{O^2}$Physics, the new analysis framework of ALICE, has been implemented. In a second step, a machine-learning model based on boosted decision trees for the separation of signal and background candidates has been developed and tested on the currently available data recorded by ALICE in Run 3.