Disentangling quark and gluon jets with normalizing flows

The isolation of pure samples quark and gluon jets is key interest at hadron colliders. Recent work has employed topic modeling to disentangle the underlying distributions in mixed obtained from experiments. However, current implementations do not scale high-dimensional observables as they rely on binning data. In this we introduce TopicFlow, a method based normalizing flows learn jet datasets. These networks are performant histogram-based approach, but since unbinned, efficient even high dimension. models can also be oversampled alleviate statistical limitations histograms. As an example use case, demonstrate how our improve calibration accuracy classifier. Finally, discuss flow likelihoods used perform outlier-robust quark/gluon classification.