| Abstract [eng] |
The aim for the master's thesis project is to develop a particle physics data generation solution that could assist CERN organization science researchers in performing high energy particle physics research work. CERN uses Monte Carlo method based data generators for high energy particle data generation. The generators use too much computer memory and processor time for data generation computations. In the master degree project classical machine learning and deep learning methods will be researched. After the performed analysis a machine learning model will be applied that uses less computer memory and processor time for particle physics data generation. The generated data will be calculated without any critical statistical flaws and based on the statistical qualities that are found in the Monte Carlo method generated dataset. The used machine learning method data generation results will be analyzed with invariant mass calculation, transverse momentum, Frechet Inception and Frobenius norm metrics. As well, the studied machine learning model computation time and memory usage will be calculated and studied to see if the applied machine learning model uses computer resources effectively. |
