Modeling the K+Σ- photoproduction in an isobar approach, using statistical learning techniques

by Dr Dimitrios Petrellis (Nuclear Physics Institute, Czech Academy of Sciences)

Tuesday 14 Jun 2022, 16:00 → 17:00 Europe/Athens

The K+Σ- photoproduction off a neutron target is examined in the framework of an isobar model. It is a phenomenological approach, where the relevant degrees-of-freedom are hadrons and their excited states. The reaction amplitude is constructed in terms of tree-level Feynman diagrams, while the coupling constants and form factors are parameters, which are fitted to experimental data and represent only effective values.  

A major challenge in this approach comes from the fact that from the numerous resonances that appear in the energy region of interest (1-2 GeV), it is not a priori known which ones participate in the reaction and therefore which ones to include in our model. We propose the use of regularization, a machine learning technique, commonly used in models with a large number of parameters, in order to prevent overfitting the data. Combined with certain information theoretic criteria, this procedure may lead to an optimal selection of the resonances involved in the reaction mechanism.

Videoconference via  https://us02web.zoom.us/j/89721925790

Dimitris Petrellis_14_06_2022.mp4