Encoding large scale cosmological structure with Generative Adversarial Networks

Abstract

Recently a type of neural networks called Generative Adversarial Networks (GANs) has been proposed as a solution for fast generation of simulation-like datasets, in an attempt to bypass heavy computations and expensive cosmological simulations to run in terms of time and computing power. In the present work, we build and train a GAN to look further into the strengths and limitations of such an approach. We then propose a novel method in which we make use of a trained GAN to construct a simple autoencoder (AE) as a first step towards building a predictive model. Both the GAN and AE are trained on images issued from two types of N-body simulations, namely 2D and 3D simulations. We find that the GAN successfully generates new images that are statistically consistent with the images it was trained on. We then show that the AE manages to efficiently extract information from simulation images, satisfyingly inferring the latent encoding of the GAN to generate an image with similar large scale structures.

arXiv commentsubmitted for publication in A&A, 15 pages

References in corpus (32)

• The cosmological simulation code GADGET-2
• Simulating the joint evolution of quasars, galaxies and their large-scale distribution
• Introducing the Illustris Project: Simulating the coevolution of dark and visible matter in the Universe
• Simulating Galaxy Formation with the IllustrisTNG Model
• GADGET: A code for collisionless and gasdynamical cosmological simulations
• How Filaments are Woven into the Cosmic Web
• The Sunyaev-Zel'dovich Effect
• A Method for Weak Lensing Observations
• The Horizon-AGN simulation: morphological diversity of galaxies promoted by AGN feedback
• Solving Large Scale Structure in Ten Easy Steps with COLA
• Adversarial Autoencoders
• Learning Particle Physics by Example: Location-Aware Generative Adversarial Networks for Physics Synthesis
• Detecting filamentary pattern in the cosmic web: a catalogue of filaments for the SDSS
• A Dynamical Classification of the Cosmic Web
• PINOCCHIO: pinpointing orbit-crossing collapsed hierarchical objects in a linear density field
• Generating Images with Perceptual Similarity Metrics based on Deep Networks
• Adversarial Video Generation on Complex Datasets
• Cosmological Structure Formation with Augmented Lagrangian Perturbation Theory
• Fast cosmic web simulations with generative adversarial networks
• Characterizing the non-linear growth of large-scale structure in the Universe
• Painting with baryons: augmenting N-body simulations with gas using deep generative models
• Learning to Denoise Astronomical Images with U-nets
• Detecting outliers in astronomical images with deep generative networks
• Fast simulation of muons produced at the SHiP experiment using Generative Adversarial Networks
• HIGAN: Cosmic Neutral Hydrogen with Generative Adversarial Networks
• Nonlinear 3D Cosmic Web Simulation with Heavy-Tailed Generative Adversarial Networks
• Inpainting Galactic Foreground Intensity and Polarization maps using Convolutional Neural Network
• A black box for dark sector physics: Predicting dark matter annihilation feedback with conditional GANs
• Deep learning for Sunyaev-Zel'dovich detection in Planck
• Exploring galaxy evolution with generative models
• Smooth stochastic density field reconstruction
• Inpainting via Generative Adversarial Networks for CMB data analysis

Cited by in corpus (10)

• Large-scale dark matter simulations
• MillimeterDL: Deep Learning Simulations of the Microwave Sky
• Investigating Cosmological GAN Emulators Using Latent Space Interpolation
• Survey2Survey: A deep learning generative model approach for cross-survey image mapping
• A search for dark matter among Fermi-LAT unidentified sources with systematic features in Machine Learning
• Stochastic Super-resolution of Cosmological Simulations with Denoising Diffusion Models
• Reionisation time fields reconstruction from 21 cm signal maps
• Reconstructing the mid-infrared spectra of galaxies using ultraviolet to submillimeter photometry and Deep Generative Networks
• Predicting large scale cosmological structure evolution with generative adversarial network-based autoencoders
• Variational autoencoder for generating realistic $N$-body simulations for dark matter halos