New graph-neural-network flavor tagger for Belle II and measurement of sin 2φ1 in B0 →J/ψ K S0 decays

We present GFlaT, a new algorithm that uses a graph-neural-network to determine the flavor of neutral B mesons produced in ϒð4SÞ decays. It improves previous algorithms by using the information from all charged final-state particles and the relations between them. We evaluate its performance using B decays to flavor-specific hadronic final states reconstructed in a 362 fb−1 sample of electron-positron collisions collected at the ϒð4SÞ resonance with the Belle II detector at the SuperKEKB collider. We achieve an effective tagging efficiency of ð37.40 0.43 0.36%Þ, where the first uncertainty is statistical and the second systematic, which is 18% better than the previous Belle II algorithm. Demonstrating the algorithm, we use B0 → J=ψK0 S decays to measure the mixing-induced and direct CP violation parameters, S ¼ ð0.724 0.035 0.009Þ and C ¼ ð−0.035 0.026 0.029Þ