Nikolai in der Wiesche - Back on Track - Neural Networks for Particle Tracking

Thursday 17 Apr 2025, 10:30 → 12:30 Europe/Stockholm

Beurling room (10238)

From the first bubble chambers to the latest silicon and gas detectors, the reconstruction of the tracks of charged particles in a magnetic field is the foundation of countless experiments in particle and hadron physics. Improving the methods used therefore offers a wide range of possibilities to increase the quantity and quality of the data used for subsequent analysis. This fact, as well as recent advances in neural network based tracking algorithms for high energy colliders, has prompted us to explore the merits of deep learning tracking algorithms for the lower energy regime. In this talk, I will present the current state of a tracking pipeline employing a graph neural network trained and tested on events simulated in the straw tube tracker of the upcoming PANDA experiment. A previous study using this pipeline has already yielded promising results for reconstructing low-momentum tracks and tracks from displaced vertices. The present work aims at further refining this approach and testing it for additional physics channels. First, I will present the performance of the pipeline on a clean sample containing only events with uniformly distributed (anti-)muons. I will then review current progress in tracking the decay products of double-strange cascade hyperons produced in proton antiproton annihilation. This is of particular scientific interest as hyperon processes are a promising probe of CP violation, but are technically challenging to study due to their long lifetimes and sequential decays resulting in multiple displaced vertices.