SARM: Sparse Autoregressive Model for Scalable Generation of Sparse Images in Particle Physics

Abstract

Generation of simulated data is essential for data analysis in particle physics, but current Monte Carlo methods are very computationally expensive. Deep-learning-based generative models have successfully generated simulated data at lower cost, but struggle when the data are very sparse. We introduce a novel deep sparse autoregressive model (SARM) that explicitly learns the sparseness of the data with a tractable likelihood, making it more stable and interpretable when compared to Generative Adversarial Networks (GANs) and other methods. In two case studies, we compare SARM to a GAN model and a non-sparse autoregressive model. As a quantitative measure of performance, we compute the Wasserstein distance ($W_p$) between the distributions of physical quantities calculated on the generated images and on the training images. In the first study, featuring images of jets in which 90% of the pixels are zero-valued, SARM produces images with $W_p$ scores that are 24-52% better than the scores obtained with other state-of-the-art generative models. In the second study, on calorimeter images in the vicinity of muons where 98% of the pixels are zero-valued, SARM produces images with $W_p$ scores that are 66-68% better. Similar observations made with other metrics confirm the usefulness of SARM for sparse data in particle physics. Original data and software will be made available upon acceptance of the manuscript from the UCI Machine Learning in Physics web portal at: http://mlphysics.ics.uci.edu/.

arXiv commentchanged templates

References in corpus (34)

• Adam: A Method for Stochastic Optimization
• Auto-Encoding Variational Bayes
• PYTHIA 6.4 Physics and Manual
• PYTHIA 6.2 Physics and Manual
• The automated computation of tree-level and next-to-leading order differential cross sections, and their matching to parton shower simulations
• Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks
• Deep Residual Learning for Image Recognition
• Large Scale GAN Training for High Fidelity Natural Image Synthesis
• DELPHES 3, A modular framework for fast simulation of a generic collider experiment
• Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks
• The ATLAS Simulation Infrastructure
• Searching for Exotic Particles in High-Energy Physics with Deep Learning
• CaloGAN: Simulating 3D High Energy Particle Showers in Multi-Layer Electromagnetic Calorimeters with Generative Adversarial Networks
• Towards Principled Methods for Training Generative Adversarial Networks
• MADE: Masked Autoencoder for Distribution Estimation
• Learning Particle Physics by Example: Location-Aware Generative Adversarial Networks for Physics Synthesis
• Deep learning in color: towards automated quark/gluon jet discrimination
• Jet-Images -- Deep Learning Edition
• Jet Substructure Classification in High-Energy Physics with Deep Neural Networks
• Jet Flavor Classification in High-Energy Physics with Deep Neural Networks
• Jet-Images: Computer Vision Inspired Techniques for Jet Tagging
• Parton Shower Uncertainties in Jet Substructure Analyses with Deep Neural Networks
• Decorrelated Jet Substructure Tagging using Adversarial Neural Networks
• CosmoGAN: creating high-fidelity weak lensing convergence maps using Generative Adversarial Networks
• Regressive and generative neural networks for scalar field theory
• RNADE: The real-valued neural autoregressive density-estimator
• Fast and accurate simulation of particle detectors using generative adversarial networks
• PixelCNN++: Improving the PixelCNN with Discretized Logistic Mixture Likelihood and Other Modifications
• Image-based model parameter optimization using Model-Assisted Generative Adversarial Networks
• Deep Neural Networks for Energy and Position Reconstruction in EXO-200
• Event Generation and Statistical Sampling for Physics with Deep Generative Models and a Density Information Buffer
• Improved Energy Reconstruction in NOvA with Regression Convolutional Neural Networks
• LHC analysis-specific datasets with Generative Adversarial Networks
• Encoding Invariances in Deep Generative Models

Cited by in corpus (10)

• Quasi Anomalous Knowledge: Searching for new physics with embedded knowledge
• Learning to Simulate High Energy Particle Collisions from Unlabeled Data
• Towards a Deep Learning Model for Hadronization
• Deep Generative Models for Detector Signature Simulation: A Taxonomic Review
• Graph Generative Adversarial Networks for Sparse Data Generation in High Energy Physics
• Generalizing to new geometries with Geometry-Aware Autoregressive Models (GAAMs) for fast calorimeter simulation
• Particle Cloud Generation with Message Passing Generative Adversarial Networks
• TopicFlow: Disentangling quark and gluon jets with normalizing flows
• Deep-learning-based decomposition of overlapping-sparse images: application at the vertex of neutrino interactions
• ParaFlow: fast calorimeter simulations parameterized in upstream material configurations