Learning Latent Jet Structure

We summarize our recent work on how to infer jet formation processes directly from substructure data using generative statistical models. recount in detail cast observables’ measurements terms of Bayesian mixed membership models, particular Latent Dirichlet Allocation. Using a sample QCD and boosted tt¯ events focusing the primary Lund plane observable basis for event measurements, we show educated priors latent distributions allows underlying physical semi-supervised way.