HIGAN: Cosmic Neutral Hydrogen with Generative Adversarial Networks

Abstract

One of the most promising ways to observe the Universe is by detecting the 21cm emission from cosmic neutral hydrogen (HI) through radio-telescopes. Those observations can shed light on fundamental astrophysical questions only if accurate theoretical predictions are available. In order to maximize the scientific return of these surveys, those predictions need to include different observables and be precise on non-linear scales. Currently, one of the best ways to achieve this is via cosmological hydrodynamic simulations; however, the computational cost of these simulations is high -- tens of millions of CPU hours. In this work, we use Wasserstein Generative Adversarial Networks (WGANs) to generate new high-resolution ($35h^{-1}\mathrm{kpc}$) 3D realizations of cosmic HI at $z=5$. We do so by sampling from a 100-dimension manifold, learned by the generator, that characterizes the fully non-linear abundance and clustering of cosmic HI from the state-of-the-art simulation IllustrisTNG. We show that different statistical properties of the produced samples -- 1D PDF, power spectrum, bispectrum, and void size function -- match very well those of IllustrisTNG, and outperform state-of-the-art models such as Halo Occupation Distributions (HODs). Our WGAN samples reproduce the abundance of HI across 9 orders of magnitude, from the Ly$\alpha$ forest to Damped Lyman Absorbers. WGAN can produce new samples orders of magnitude faster than hydrodynamic simulations.

arXiv comment9 pages, 8 figures

References in corpus (9)

• Improved Training of Wasserstein GANs
• MMD GAN: Towards Deeper Understanding of Moment Matching Network
• Learning to Predict the Cosmological Structure Formation
• Distinguishing standard and modified gravity cosmologies with machine learning
• Inflation and Early Dark Energy with a Stage II Hydrogen Intensity Mapping Experiment
• Painting halos from cosmic density fields of dark matter with physically motivated neural networks
• Denoising Weak Lensing Mass Maps with Deep Learning
• Intensity mapping with neutral hydrogen and the Hidden Valley simulations
• Cosmological model discrimination with Deep Learning

Cited by in corpus (21)

• Solving inverse-PDE problems with physics-aware neural networks
• Super-resolution emulator of cosmological simulations using deep physical models
• Painting halos from cosmic density fields of dark matter with physically motivated neural networks
• Nonlinear 3D Cosmic Web Simulation with Heavy-Tailed Generative Adversarial Networks
• From EMBER to FIRE: predicting high resolution baryon fields from dark matter simulations with Deep Learning
• Deep learning for intensity mapping observations: Component extraction
• Encoding large scale cosmological structure with Generative Adversarial Networks
• A unified framework for 21cm tomography sample generation and parameter inference with Progressively Growing GANs
• Large-scale structures in the $\Lambda$CDM Universe: network analysis and machine learning
• Fast, high-fidelity Lyman $\alpha$ forests with convolutional neural networks
• Modeling assembly bias with machine learning and symbolic regression
• HyPhy: Deep Generative Conditional Posterior Mapping of Hydrodynamical Physics
• The Sejong Suite: Cosmological Hydrodynamical Simulations with Massive Neutrinos, Dark Radiation, and Warm Dark Matter
• Deep Forest: Neural Network reconstruction of the Lyman-alpha forest
• Understanding the Impact of Semi-numeric Reionization Models when using CNNs
• Predicting 21cm-line map from Lyman $\alpha$ emitter distribution with Generative Adversarial Networks
• Uncertainties associated with GAN-generated datasets in high energy physics
• Modeling Post-Reionization HI Distributions in Fuzzy Dark Matter Cosmologies Using Conditional Normalizing Flows
• Stochastic Super-resolution of Cosmological Simulations with Denoising Diffusion Models
• Deconfusing intensity maps with neural networks
• Neural Symplectic Integrator with Hamiltonian Inductive Bias for the Gravitational $N$-body Problem