The field of machine learning and deep learning has made tremendous progress in recent years. Machine Learning techniques are used in particle physics for many decades and are today used in almost any physics analysis.

A rather new and exiting sub branch of machine learning are generative adversarial networks (GANs). Our group is actively studying the possibilty to at least partly replace the computationally very expensive detailed simulation of particle showers in highly granular calorimeters with Geant4 by the application of novel generative networks architectures.

First studies using a Bounded-Information-Bottleneck Autoencoder (BIB-AE) show very promising first results in generating high fidelity photon showers in a highly granular calorimter with high speed.