Getting familiar with jet flavour tagging

Before we start directly with the project, familiarise yourself with the techniques and terminology used in ATLAS jet flavour tagging.

Task 1

Read the ATLAS Run-2 flavour tagging paper "Eur. Phys. J. C 79 (2019): ATLAS b-jet identification performance and efficiency measurement with tt events in pp collisions at √s = 13 TeV" .

Focus on the sections 1–4, 10.5 and 11 which describe the algorithms and their input features. The other sections discuss how the algorithms, which were trained and evaluated in MC simulation, perform in observed data and how to correct for the residual differences between data and simulation using a calibration procedure.

Note down any questions you have. You should have plenty, because it is a summary paper and a lot of details are just hidden in the references. You don't have to track down the references. After reading, you should be able to answer these questions:

Questions:

• What are the distinctive features of b-jets and how are they used to distinguish b-jets from jets initiated by c-quarks and light-flavour quarks (u, d, s) and gluons?

• Which parts of the ATLAS detector are essential to the identification of b-jets?

• How do we know in our MC simulation that a jet originates from a b-quark? How do we know that it is originating from a c-quark?

• Which low-level algorithms are used in ATLAS to identify observables sensitive to the presence of b-hadrons and their displaced decay at a secondary vertex?

• How are the low-level algorithm observables combined into a single discriminant?

Task 2

Complete the first two exercises (all parts of exercise 1 and exercise 2) of the Puma plotting tutorial.

You can use the jupyter environment at NAF (see Computing resources section) for the tutorial. Remember to select the ftag python Kernel to use the miniconda environment which has puma and matplotlib installed.

You can find the tutorial in the section "Software tutorials".

If you make fast progress, you can tackle also the exercises 3 and 4 in the tutorial. If you completed those, try to explore the variables in the dataset. Try plotting not only the jet "pt" but also the variables you can find inside the h5 file for the jets dataset.