Heavy flavor identification at CMS with deep neural networks

At the Large Hadron Collider, the identification of jets originating from heavy flavour quarks (b or c-tagging) is important for searches for new physics and for measurements of standard model processes. A variety of b-tagging algorithms has been developed by CMS to select b-quark jets based on variables such as the impact parameters of the charged-particle tracks, the properties of reconstructed decay vertices, and the presence or absence of a lepton, or combinations thereof. These algorithms heavily rely on machine learning tools and are thus natural candidates for advanced tools like deep neural networks. A new algorithm, DeepCSV, uses a deep neural network. The input is the same set of observables used by the existing CSVv2 b-tagger, with the extension that it uses information of more tracks. Also, the training strategy was adapted and about 50 million jets are used for the training of the deep neural network. The new DeepCSV algorithm outperforms the CSVv2 tagger, with an absolute b-tagging efficiency improvement of about 4\% for a misidentification probability for light-flavour jets of 1\%. In addition, DeepCSV is a multiclassifier simultaneously trained for c-tagging. For c-tagging DeepCSV outperforms the other taggers in CMS.