Usage of Generative NNs for the Fast Simulation of Charged Particle Interactions With Matter

The precise modeling of the charged particles interaction is the most important task for the advancement of nuclear and particle physics. One of the ways to study such interactions is to use detectors at Large Hadron Collider (LHC) to select interesting particles collisions and analyze them. While the next big run of Large Hadron Collider is planned to 2021 more and more data about collisions are requested by scientists to build more precise statistical models for particle interactions. The classical approach to generate more data is to use Markov chain methods which have a number of drawbacks like time complexity. In this work, we are going to present a better yet unexplored method for sampling detector responses. Our plan is to show the superiority of such a method called Cram\'er GAN over MCMC methods in terms of performance