Santiago Tanco: Exploring unsupervised top tagging using Bayesian inference

Santiago Tanco: Exploring unsupervised top tagging using Bayesian inference

Identifying hadronically decaying top-quark jets in a sample of jets is an important task in many LHC searches. Although there are outstanding supervised algorithms, their construction and expected performance rely on Monte Carlo simulations. After introducing the general framework of Bayesian inference over mixture models, I will present some results on the development of two simple unsupervised top tagging algorithms based on these techniques.