Speaker
Savelii Omelyanchuk
(JINR)
Description
This work is dedicated to the development of deep learning methods for particle track classification. It focuses on the graph neural network (GNN) architecture for classifying tracks by events in each time slice at the SPD experiment. The work presents a novel approach to track sorting, an analysis of training dynamics, and model testing under various conditions. The model is implemented and trained using modern deep machine learning tools that enable parallel tensor computations.
