Nonlinear 3D Cosmic Web Simulation with Heavy-Tailed Generative Adversarial Networks

Abstract

Fast and accurate simulations of the non-linear evolution of the cosmic density field are a major component of many cosmological analyses, but the computational time and storage required to run them can be exceedingly large. For this reason, we use generative adversarial networks (GANs) to learn a compressed representation of the 3D matter density field that is fast and easy to sample, and for the first time show that GANs are capable of generating samples at the level of accuracy of other conventional methods. Using sub-volumes from a suite of GADGET-2 N-body simulations, we demonstrate that a deep-convolutional GAN can generate samples that capture both large- and small-scale features of the matter density field, as validated through a variety of n-point statistics. The use of a data scaling that preserves high-density features and a heavy-tailed latent space prior allow us to obtain state of the art results for fast 3D cosmic web generation. In particular, the mean power spectra from generated samples agree to within 5% up to k=3 and within 10% for k<5 when compared with N-body simulations, and similar accuracy is obtained for a variety of bispectra. By modeling the latent space with a heavy-tailed prior rather than a standard Gaussian, we better capture sample variance in the high-density voxel PDF and reduce errors in power spectrum and bispectrum covariance on all scales. Furthermore, we show that a conditional GAN can smoothly interpolate between samples conditioned on redshift. Deep generative models, such as the ones described in this work, provide great promise as fast, low-memory, high-fidelity forward models of large-scale structure.

arXiv comment19 pages, 17 figures. v3: Reflects changes in version published in PRD

References in corpus (37)

• Array Programming with NumPy
• Conditional Generative Adversarial Nets
• Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks
• GANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibrium
• Large Scale GAN Training for High Fidelity Natural Image Synthesis
• Learning a Probabilistic Latent Space of Object Shapes via 3D Generative-Adversarial Modeling
• How Filaments are Woven into the Cosmic Web
• Solving Large Scale Structure in Ten Easy Steps with COLA
• The clustering of galaxies in the SDSS-III Baryon Oscillation Spectroscopic Survey: a large sample of mock galaxy catalogues
• Learning to Predict the Cosmological Structure Formation
• FastPM: a new scheme for fast simulations of dark matter and halos
• Modelling Baryon Acoustic Oscillations with Perturbation Theory and Stochastic Halo Biasing
• Gauge Equivariant Convolutional Networks and the Icosahedral CNN
• An accurate tool for the fast generation of dark matter halo catalogs
• The Dark Universe
• CosmoGAN: creating high-fidelity weak lensing convergence maps using Generative Adversarial Networks
• The clustering of galaxies in the SDSS-III Baryon Oscillation Spectroscopic Survey: cosmological constraints from the full shape of the clustering wedges
• ICE-COLA: Towards fast and accurate synthetic galaxy catalogues optimizing a quasi $N$-body method
• The mass-Peak Patch algorithm for fast generation of deep all-sky dark matter halo catalogues and its N-Body validation
• Fast cosmic web simulations with generative adversarial networks
• Cosmological Reconstruction From Galaxy Light: Neural Network Based Light-Matter Connection
• Painting with baryons: augmenting N-body simulations with gas using deep generative models
• Super-resolution emulator of cosmological simulations using deep physical models
• HALOGEN: A tool for fast generation of mock halo catalogues
• A volumetric deep Convolutional Neural Network for simulation of mock dark matter halo catalogues
• Painting halos from cosmic density fields of dark matter with physically motivated neural networks
• Cosmic homogeneity: a spectroscopic and model-independent measurement
• HIGAN: Cosmic Neutral Hydrogen with Generative Adversarial Networks
• Predicting dark matter halo formation in N-body simulations with deep regression networks
• Non-Linear Matter Power Spectrum Covariance Matrix Errors and Cosmological Parameter Uncertainties
• Powerbox: A Python package for creating structured fields with isotropic power spectra
• Testing homogeneity in the Sloan Digital Sky Survey Data Release Twelve with Shannon entropy
• A black box for dark sector physics: Predicting dark matter annihilation feedback with conditional GANs
• Uncertainty Quantification with Generative Models
• Perfectly parallel cosmological simulations using spatial comoving Lagrangian acceleration
• Connecting Dualities and Machine Learning
• Non-Parametric Priors For Generative Adversarial Networks

Cited by in corpus (14)

• Large-scale dark matter simulations
• The DAWES review 10: The impact of deep learning for the analysis of galaxy surveys
• Super-resolution emulator of cosmological simulations using deep physical models
• Removing Astrophysics in 21 cm maps with Neural Networks
• From EMBER to FIRE: predicting high resolution baryon fields from dark matter simulations with Deep Learning
• Encoding large scale cosmological structure with Generative Adversarial Networks
• Fast and realistic large-scale structure from machine-learning-augmented random field simulations
• A Machine Learning Approach to Correct for Mass Resolution Effects in Simulated Halo Clustering Statistics
• Psi-GAN: A power-spectrum-informed generative adversarial network for the emulation of large-scale structure maps across cosmologies and redshifts
• Machine Learning the Fates of Dark Matter Subhalos: A Fuzzy Crystal Ball
• Stochastic Super-resolution of Cosmological Simulations with Denoising Diffusion Models
• Pareto GAN: Extending the Representational Power of GANs to Heavy-Tailed Distributions
• nuGAN: Generative Adversarial Emulator for Cosmic Web with Neutrinos
• On some theoretical limitations of Generative Adversarial Networks