The Belle II flavor tagger

Belle II is a particle-physics experiment at the intensity frontier focused on probing non Standard Model physics through precision measurements of quark-flavor and τ-lepton dynamics. Determining the flavor of neutral B mesons, i.e. their quark composition, is a crucial task which is addressed using flavor tagging algorithms. Due to the novel high-luminosity conditions and the increased beam backgrounds at Belle II, an improved flavor tagging algorithm had to be developed to ensure the success of the Belle II physics program. The new Belle II flavor tagger exploits the flavor-specific signatures of B 0 decays employing boosted decision trees and neural networks. It identifies B 0-decay products providing flavor-specific signatures and combines the information from all possible signatures into a final output. The algorithm has been validated by comparing its performance on simulated events with its performance on collision events collected by the predecessor experiment Belle. To explore the advantages of state-of-the-art deep-learning techniques, the Belle II collaboration developed a deep-learning-based flavor tagger. This algorithm tags the flavor of B 0 mesons without identifying flavor specific signatures using a deep-learning neural network. The validation on Belle data of this algorithm is currently ongoing.