Data-driven jet classification with Latent Dirichlet Allocation
In this talk I will discuss how ideas borrowed from the machine learning (ML) communities can be used to construct data-driven and unsupervised classifiers for jets. These borrowed ideas come from a branch of ML called "topic modelling", where the Latent Dirichlet Allocation (LDA) algorithm that we use was first developed. Initially constructed to study topics occurring within a set of text documents, the algorithm has since been applied to several areas of scientific study.
For help, please contact Webmaster.