Dortmund 2021 – wissenschaftliches Programm

Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

T: Fachverband Teilchenphysik

T 57: Calorimeters I

T 57.1: Vortrag

Mittwoch, 17. März 2021, 16:00–16:15, Tg

Even π-er: High Fidelity Simulation of Pion Showers with High Speed — •Sascha Diefenbacher — Institut für Experimentalphysik, Universität Hamburg, Deutschland

Simulations of particle collisions play an irreplaceable part in particle physics analyses. A significant part of the time required for these simulations has to be dedicated to modeling particle showers. As collider luminosities continue to increase, the demand for simulated data, and thereby the required simulation time, increases with it. It is therefore imperative that we find ways to speed up these costly simulations. In previous works we managed to show that generative machine learning models can be used to simulate electromagnetic photon showers significantly faster than classical simulations methods. Building on this success we now attempt to extend our generative setups to simulate pion showers. Compared to the electromagnetic showers these pion showers feature significantly more variance in their structure, which presents an additional challenge for the generative models. We present ongoing results using a Generative Adversarial Network (GAN), a Wasserstein GAN (WGAN) and a Bounded Information Bottleneck Autoencoder (BIB-AE).