AI-supported analysis of hadron spectroscopy data

Ambito della ricerca

Fisica sperimentale delle interazioni fondamentali

Abstract

Despite a tremendous effort in the last 50 years, strong interactions remain poorly understood and fundamental questions such as where the mass and the spin of the proton come from, still need to be answered.
A rich phenomenology of the hadron spectrum is available from data collected at colliders (LHCb, BESIII, Belle2, ...) and fixed target experiments (GLUEX, JLAB, MAINZ, ...) but the interpretation lacks a coherent description and full understanding. Part of the problem is related to the difficulties in extracting information from the multidimensional phase space that describes exclusive hadron-production channels. AI-based algorithms, such as generative machine learning models (ML GANs) provide a novel and promising tool to analyze and reproduce experimental data, unfold the detector effects and extract the production amplitudes. In this thesis the candidate will learn the fundamentals of generative ML models and their application to hadron physics. The work will be performed in collaboration with experimental and theoretical physicists of Jefferson Lab (US) and data scientists from Old Dominion University (US) and Roma Tor Vergata University.  During the thesis, a stage in US is foreseen.