Uncertainties associated with GAN-generated datasets in high energy physics

Abstract

Recently, Generative Adversarial Networks (GANs) trained on samples of traditionally simulated collider events have been proposed as a way of generating larger simulated datasets at a reduced computational cost. In this paper we point out that data generated by a GAN cannot statistically be better than the data it was trained on, and critically examine the applicability of GANs in various situations, including a) for replacing the entire Monte Carlo pipeline or parts of it, and b) to produce datasets for usage in highly sensitive analyses or sub-optimal ones. We present our arguments using information theoretic demonstrations, a toy example, as well as in the form of a formal statement, and identify some potential valid uses of GANs in collider simulations.

28 pages, 11 figures. Revised submission to SciPost

References in corpus (24)

• Conditional Generative Adversarial Nets
• Event Generation with Normalizing Flows
• How to GAN LHC Events
• Calorimetry with Deep Learning: Particle Simulation and Reconstruction for Collider Physics
• i-flow: High-dimensional Integration and Sampling with Normalizing Flows
• Exploring phase space with Neural Importance Sampling
• Invertible Networks or Partons to Detector and Back Again
• How to GAN away Detector Effects
• Simulation of electron-proton scattering events by a Feature-Augmented and Transformed Generative Adversarial Network (FAT-GAN)
• GANplifying Event Samples
• A Neural Resampler for Monte Carlo Reweighting with Preserved Uncertainties
• Fast simulation of muons produced at the SHiP experiment using Generative Adversarial Networks
• Lund jet images from generative and cycle-consistent adversarial networks
• From Lagrangians to Events: Computer Tutorial at the MC4BSM-2012 Workshop
• Noise reduction for weak lensing mass mapping: An application of generative adversarial networks to Subaru Hyper Suprime-Cam first-year data
• HIGAN: Cosmic Neutral Hydrogen with Generative Adversarial Networks
• Efficient Monte Carlo Integration Using Boosted Decision Trees and Generative Deep Neural Networks
• LHC analysis-specific datasets with Generative Adversarial Networks
• Particle Generative Adversarial Networks for full-event simulation at the LHC and their application to pileup description
• How to GAN Event Subtraction
• A black box for dark sector physics: Predicting dark matter annihilation feedback with conditional GANs
• Computational challenges for MC event generation
• CMB-GAN: Fast Simulations of Cosmic Microwave background anisotropy maps using Deep Learning
• OASIS: Optimal Analysis-Specific Importance Sampling for event generation

Cited by in corpus (4)

• i-flow: High-dimensional Integration and Sampling with Normalizing Flows
• Accelerating Monte Carlo event generation -- rejection sampling using neural network event-weight estimates
• A survey of machine learning-based physics event generation
• Synthesis of pulses from particle detectors with a Generative Adversarial Network (GAN)