Deutsches Elektronen-Synchrotron DESY

Helmholtz-Partner:

CASUS
Helmholtz-Zentrum Dresden-Rossendorf e.V. HZDR

Simulations of fundamental processes in physics that are relevant for high-energy physics and materials science can often be achieved only with computational techniques that scale unfavorably with system parameters and have a high computational overhead. Fueled by the universal approximation theorem, neural networks have great potential to accelerate our conventional simulation techniques. The multicenter team from DESY and CASUS/HZDR tackles the computational challenge of constructing accurate neural networks for fundamental physics applications.

High-fidelity simulated events that are needed to train neural networks are often sparse and very costly to generate. The SynRap team investigates the generation of synthetic data using surrogate models in order to produce a significantly larger amount of training data. If the team finds that these data reproduce the original simulated events accurately, the synthetic data can be used in a second step for an improved training of the neural networks. Methodologically, this project is at the forefront of active research. If successful, this project will significantly boost computational efficiency, enabling large-scale simulations currently unattainable with standard methods.

A singular feature of the Synrap collaboration is the level of abstraction. The team will develop a unified surrogate model framework that will be used to tackle common challenges in two seemingly different research areas – high-energy physics (HEP) and high energy-density (HED) phenomena. While scales and physics in both fields are very different, methodological challenges are similar.

Coordination at DESY:

Deutsches Elektronen-Synchrotron
Dr. Isabell Melzer-Pellmann
Notkestr. 85
D-22607 Hamburg

Email: isabell.melzer@desy.de
Phone: +49 40 8998 2489

Helmholtz-Partner:
Center for Advanced Systems Understanding CASUS
Helmholtz-Zentrum Dresden-Rossendorf e.V. HZDR

Dr. Attila Cangi
Email: a.cangi@hzdr.de
Phone: +49 3581 3752352