Thesis BELLE2-MTHESIS-2022-039

Development of a Graph Neural Network based Tagging Approach for Semi-Inclusive Tagging and Rest of Event Clean Up for the Belle II Experiment

Prof. Dr. Florian Bernlochner ; Prof. Dr. Jochen Dingfelder

University of Bonn Bonn

Abstract: This thesis focuses on the development of software tools which shall improve the particle reconstruction at Belle II. The tools are based on a graph neural network, called the DSIT model, and aim to categorize the final state particles of a decay event in order to improve the tagging. This is done as a further development of an already existing semi-inclusive tagging method, the further development being named deep semi-inclusive tagging in this thesis. The main goal is to improve purity and efficiency of the tagging. Later the DSIT model’s capability to clean up the rest of event after the signal 𝐵 meson was already reconstructed gets investigated.

Note: Presented on 31 03 2022
Note: MSc

The record appears in these collections:
Books, Theses & Reports > Theses > Masters Theses

 Record created 2022-11-03, last modified 2022-11-08

Download fulltext

Rate this document:

Rate this document:
(Not yet reviewed)