Speaker
Description
The Z-pole operation of CEPC will provide a large data sample for precision electroweak measurements. We present measurements of the forward-backward asymmetries $A_{FB}$ in Z decays using the CEPC reference detector simulation, and further derive the effective weak mixing angle $\sin^2\theta_W^{\mathrm{eff}}$ from these observables. The Particle Transformer deep learning architecture is employed to improve the efficiency and purity for bottom, charm, and strange quarks. Based on the resulting tagging performance, we extract the asymmetry observables and evaluate their statistical uncertainties, along with the corresponding precision of $\sin^2\theta_W^{\mathrm{eff}}$ under unpolarized beam conditions.
Incorporating beam polarization inspired by SLD, we simulate longitudinally polarized cross-sections to assess its enhancement on the weak mixing angle measurement. Beyond statistical precision, we provide both qualitative and quantitative evaluations of major systematic uncertainties with detailed assessments of their impacts on our results. Our study demonstrates that the combination of polarization and advanced AI could push the precision of these electroweak observables at CEPC to a globally leading level.
