SymmetryGAN: Symmetry Discovery with Deep Learning

Abstract

What are the symmetries of a dataset? Whereas the symmetries of an individual data element can be characterized by its invariance under various transformations, the symmetries of an ensemble of data elements are ambiguous due to Jacobian factors introduced while changing coordinates. In this paper, we provide a rigorous statistical definition of the symmetries of a dataset, which involves inertial reference densities, in analogy to inertial frames in classical mechanics. We then propose SymmetryGAN as a novel and powerful approach to automatically discover symmetries using a deep learning method based on generative adversarial networks (GANs). When applied to Gaussian examples, SymmetryGAN shows excellent empirical performance, in agreement with expectations from the analytic loss landscape. SymmetryGAN is then applied to simulated dijet events from the Large Hadron Collider (LHC) to demonstrate the potential utility of this method in high energy collider physics applications. Going beyond symmetry discovery, we consider procedures to infer the underlying symmetry group from empirical data.

19 pages, 17 figures

References in corpus (31)

• Adam: A Method for Stochastic Optimization
• PYTHIA 6.4 Physics and Manual
• PYTHIA 6.2 Physics and Manual
• TensorFlow: A system for large-scale machine learning
• The anti-k_t jet clustering algorithm
• A Brief Introduction to PYTHIA 8.1
• FastJet user manual
• Generative Adversarial Networks
• DELPHES 3, A modular framework for fast simulation of a generic collider experiment
• Dispelling the N^3 myth for the Kt jet-finder
• Variational Inference with Normalizing Flows
• Normalizing Flows: An Introduction and Review of Current Methods
• Data Augmentation Generative Adversarial Networks
• Hamiltonian Neural Networks
• Anomaly Detection for Resonant New Physics with Machine Learning
• Parameterized Machine Learning for High-Energy Physics
• Learning New Physics from a Machine
• Extending the Bump Hunt with Machine Learning
• OmniFold: A Method to Simultaneously Unfold All Observables
• Lorentz Group Equivariant Neural Network for Particle Physics
• GANplifying Event Samples
• A Neural Resampler for Monte Carlo Reweighting with Preserved Uncertainties
• Equivariant Energy Flow Networks for Jet Tagging
• DCTRGAN: Improving the Precision of Generative Models with Reweighting
• Symmetry meets AI
• A method to challenge symmetries in data with self-supervised learning
• Learning Invariances in Neural Networks
• Meta-Learning Symmetries by Reparameterization
• Particle Convolution for High Energy Physics
• Set2Graph: Learning Graphs From Sets
• Fast AutoAugment

Cited by in corpus (10)

• Machine Learning and LHC Event Generation
• Event Generators for High-Energy Physics Experiments
• Calomplification -- The Power of Generative Calorimeter Models
• Deep Generative Models for Detector Signature Simulation: A Taxonomic Review
• A method to challenge symmetries in data with self-supervised learning
• Discovering Sparse Representations of Lie Groups with Machine Learning
• Data-driven discovery of self-similarity using neural networks
• Extracting self-similarity from data
• Automated detection of symmetry-protected subspaces in quantum simulations
• SymmetryLens: Unsupervised Symmetry Learning via Locality and Density Preservation