% IMPORTANT: The following is UTF-8 encoded.  This means that in the presence
% of non-ASCII characters, it will not work with BibTeX 0.99 or older.
% Instead, you should use an up-to-date BibTeX implementation like “bibtex8” or
% “biber”.

@INPROCEEDINGS{Scham:905718,
      author       = {Scham, Moritz and Bhattacharya, Soham and Borras, Kerstin
                      and Fedorko, Wojtek and Jitsev, Jenia and Katzy, Judith and
                      Krücker, Dirk},
      title        = {{D}e{G}e{S}im and the {H}igh {G}ranularity {C}alorimeter
                      for the {CMS} {E}xperiment at the {L}arge {H}adron
                      {C}ollider},
      reportid     = {FZJ-2022-00943},
      year         = {2021},
      abstract     = {We present the beginning work on fast simulations for the
                      new High Granularity Calorimeter (HGCal), which will become
                      part of the CMS experiment at the Large Hadron Colider (LHC)
                      at CERN/Geneva. It will replace the current end-cap
                      calorimeters for data acquisition in 2027. The design of the
                      new calorimeter is driven by the increased particle density
                      following the High Luminosity Upgrade of the LHC. This
                      enforces a complex geometry of the active sensor and a
                      massive increase from about 7k currently to more than 3
                      million individual channels for each of the two future
                      end-cap calorimeters. Traditionally, particle physics has
                      used detailed simulations of detectors and physics. These
                      simulations are of high precision and can describe minute
                      details. The price for this precision is a large
                      computational power requirement. Since the predicted
                      computational resources will not be sufficient for the
                      traditional stepwise simulations, we will explore deep
                      learning techniques combining fast inference with generative
                      models, e.g., physics-informed GANs-VAEs hybrids for
                      simulations. In particular, Graph Neural Networks are a
                      promising candidate to describe thecomplex geometry of the
                      new detector. This project is part of the Helmholtz AI
                      funded project DeGeSim.},
      month         = {Apr},
      date          = {2021-04-14},
      organization  = {Helmholtz AI virtual conference 2021,
                       online (Germany), 14 Apr 2021 - 15 Apr
                       2021},
      subtyp        = {After Call},
      cin          = {JSC},
      cid          = {I:(DE-Juel1)JSC-20090406},
      pnm          = {5112 - Cross-Domain Algorithms, Tools, Methods Labs (ATMLs)
                      and Research Groups (POF4-511) / ZT-I-PF-5-3 - Deep
                      Generative models for fast and precise physics Simulation
                      (DeGeSim) $(2020_ZT-I-PF-5-3)$},
      pid          = {G:(DE-HGF)POF4-5112 / $G:(DE-HGF)2020_ZT-I-PF-5-3$},
      typ          = {PUB:(DE-HGF)24},
      url          = {https://juser.fz-juelich.de/record/905718},
}