Ing. Marcel Jiřina, DrSc.

Scientist

Department: Machine Learning
Phone: +420 266 05 3350
Office:
Email: (email)

Project PPDS - Particle Physics Data Separation

This project group has two members in the Institute of Computer Science AS CR, Marcel Jiřina and František Hakl. Our group and our work is conected to the D0 experiment at FERMILAB via the Faculty of Nuclear Sciences and Physical Engineering of the Czech Technical University in Prague.

Motivation

In particle physics experimeters are faced to general problem of useful data (called signal) and huge amount of unuseful data (called background) separation (filtration) with aim to suppress the background and keep the signal as much as possible. There is a standard approach of "cuts" that consists of setting suitable thresholds to individual variables and cutting a part of its statistics. We deal with problem of signal - background events separation for various particle decay processes using alternative approaches of sophisticated neural networks and methods based on fractal nature of multivariate data.