Fast simulation methods in ATLAS: from classical to generative models

¹ Cavendish Laboratory, University of Cambridge, Cambridge, United Kingdom
² Department of Physics, New York University, New York, NY, USA
³ Simons Mobility Gmbh, Munich, Bavaria, Germany
⁴ Facultè des Sciences, Département de Physique Nucléaire et Corpusculaire (DPNC), Université de Genève 24, Quai Ernest-Ansermet, CH-1211 Genève 4, Geneva, Switzerland
⁵ IJCLab, Université Paris-Saclay, CNRS/IN2P3, 91405, Orsay, France
⁶ Physics Department, University of California, Berkeley, CA, United States of America
⁷ Physics Division, Lawrence Berkeley National Laboratory, Berkeley, CA, United States of America
⁸ INFN Sezione di Milano ; Milano, Italy
⁹ Department of Physics, University of Washington, Seattle WA, United States of America

^* e-mail: heather.gray@berkeley.edu

Published online: 16 November 2020

Abstract

The ATLAS physics program relies on very large samples of Geant4 simulated events, which provide a highly detailed and accurate simulation of the ATLAS detector. However, this accuracy comes with a high price in CPU, and the sensitivity of many physics analyses is already limited by the available Monte Carlo statistics and will be even more so in the future. Therefore, sophisticated fast simulation tools have been developed. In Run 3 we aim to replace the calorimeter shower simulation for most samples with a new parametrised description of longitudinal and lateral energy deposits, including machine learning approaches, to achieve a fast and accurate description. Looking further ahead, prototypes are being developed using cutting edge machine learning approaches to learn the appropriate calorimeter response, which are expected to improve modeling of correlations within showers. Two different approaches, using Variational Auto-Encoders (VAEs) or Generative Adversarial Networks (GANs), are trained to model the shower simulation. Additional fast simulation tools will replace the inner detector simulation, as well as digitization and reconstruction algorithms, achieving up to two orders of magnitude improvement in speed. In this talk, we will describe the new tools for fast production of simulated events and an exploratory analysis of the deep learning methods.