Dear All,
we have the pleasure to invite you to the Particle Theory Seminar on Tuesday, March 28 at 12:15, which will be given by
Paweł Sznajder (NCBJ)
entitled
"Machine learning techniques in phenomenology of generalised parton distributions (GPDs)”
Abstract:
"Generalised partons distributions (GPDs) describe the internal structure of hadrons in terms of quarks and gluons. They can be used, for instance, to study so-called nucleon tomography, and to access elements of the QCD energy-momentum tensor. Reconstruction of GPDs is a particularly difficult task, as available data are sparse compared to the number of objects and dimensions one must consider. In addition, one needs to solve a difficult inverse problem and fulfil a set of not trivial theory-driven constraints. In my talk I will show how elements of machine learning can be adapted and used for these purposes. I will focus on artificial neural networks, stressing their complementarity to classic techniques, like parametric fits. Technicalities involving e.g. genetic algorithms will also be described. "
The seminar will be in a hybrid mode, local people are invited to meet in the seminar room D-2-02.
Piotr Korcyl, Jacek Wosiek, Michał Praszałowicz