A deep learning model to emulate simulations of cosmic reionization

Abstract

We present a deep learning model trained to emulate the radiative transfer during the epoch of cosmological reionization. CRADLE (Cosmological Reionization And Deep LEarning) is an autoencoder convolutional neural network that uses two-dimensional maps of the star number density and the gas density field at z=6 as inputs and that predicts 3D maps of the times of reionization ${\mathrm{t}}_{\mathrm{reion}}$ as outputs. These predicted single fields are sufficient to describe the global reionization history of the intergalactic medium in a given simulation. We trained the model on a given simulation and tested the predictions on another simulation with the same paramaters but with different initial conditions. The model is successful at predicting ${\mathrm{t}}_{\mathrm{reion}}$ maps that are in good agreement with the test simulation. We used the power spectrum of the ${\mathrm{t}}_{\mathrm{reion}}$ field as an indicator to validate our model. We show that the network predicts large scales almost perfectly but is somewhat less accurate at smaller scales. While the current model is already well-suited to get average estimates about the reionization history, we expect it can be further improved with larger samples for the training, better data pre-processing and finer tuning of hyper-parameters. Emulators of this kind could be systematically used to rapidly obtain the evolving HII regions associated with hydro-only simulations and could be seen as precursors of fully emulated physics solvers for future generations of simulations.

arXiv comment11 pages, 8 figures, accepted in MNRAS (official release 21st of october)

References in corpus (47)

• In the Beginning: The First Sources of Light and the Reionization of the Universe
• UV Luminosity Functions at redshifts z~4 to z~10: 10000 Galaxies from HST Legacy Fields
• The Growth of HII Regions During Reionization
• Cosmological Reionization by Stellar Sources
• Simulating Cosmic Reionization at Large Scales I: the Geometry of Reionization
• Stellar Population Effects on the Inferred Photon Density at Reionization
• RAMSES-RT: Radiation hydrodynamics in the cosmological context
• Simulations and Analytic Calculations of Bubble Growth During Hydrogen Reionization
• The SPHINX Cosmological Simulations of the First Billion Years: the Impact of Binary Stars on Reionization
• Multi--dimensional Cosmological Radiative Transfer with a Variable Eddington Tensor Formalism
• The History of the Baryon Budget: Cosmic Logistics in a Hierarchical Universe
• Cosmic Dawn (CoDa): the First Radiation-Hydrodynamics Simulation of Reionization and Galaxy Formation in the Local Universe
• Comparison Of Reionization Models: Radiative Transfer Simulations And Approximate, Semi-Numeric Models
• Enzo+Moray: Radiation Hydrodynamics Adaptive Mesh Refinement Simulations with Adaptive Ray Tracing
• Calibration Requirements for Detecting the 21 cm Epoch of Reionization Power Spectrum and Implications for the SKA
• Inside-out or Outside-in: The topology of reionization in the photon-starved regime suggested by Lyman-alpha forest data
• Cosmic Reionization On Computers I. Design and Calibration of Simulations
• Interpreting ALMA Observations of the ISM During the Epoch of Reionisation
• Experimental Constraints on Self-consistent Reionization Models
• Reionization simulations powered by GPUs I: the structure of the Ultraviolet radiation field
• Cosmic Reionization On Computers II. Reionization History and Its Back Reaction on Early Galaxies
• Deep learning from 21-cm tomography of the Cosmic Dawn and Reionization
• Fast Large-Scale Reionization Simulations
• Fast cosmic web simulations with generative adversarial networks
• Analysing the 21cm signal from the Epoch of Reionization with artificial neural networks
• Deep Convolutional Neural Networks as strong gravitational lens detectors
• A Semi-Analytic Model for Cosmological Reheating and Reionization Due to the Gravitational Collapse of Structure
• Deep Learning of Quasar Spectra to Discover and Characterize Damped Lya Systems
• The Epoch of Reionization
• Connecting Reionization to the Local Universe
• Machine Learning and Cosmological Simulations II: Hydrodynamical Simulations
• Identifying Reionization Sources from 21cm Maps using Convolutional Neural Networks
• EMMA: an AMR cosmological simulation code with radiative transfer
• Machine Learning and Cosmological Simulations I: Semi-Analytical Models
• How JWST can measure First Light, Reionization and Galaxy Assembly
• Impact of the reduced speed of light approximation on ionization front velocities in cosmological simulations of the epoch of reionization
• The Inhomogeneous Reionization Times of Present-day Galaxies
• The Role of Machine Learning in the Next Decade of Cosmology
• Self-shielding of hydrogen in the IGM during the epoch of reionization
• Impact of the reduced speed of light approximation on the post-overlap neutral hydrogen fraction in numerical simulations of the epoch of reionization
• On the Proper Use of the Reduced Speed of Light Approximation
• High resolution simulations of the reionization of an isolated Milky Way - M31 galaxy pair
• GAME: GAlaxy Machine learning for Emission lines
• A study of simulated histories of reionization with merger trees of HII regions
• The interstellar medium of dwarf galaxies: new insights from Machine Learning analysis of emission line spectra
• Probing individual sources during reionization and cosmic dawn using SKA HI 21-cm observations
• Local reionizations histories with merger tree of HII regions

Cited by in corpus (32)

• The data-driven future of high energy density physics
• Dimensionality Reduction of SDSS Spectra with Variational Autoencoders
• Constraining the astrophysics and cosmology from 21cm tomography using deep learning with the SKA
• deep21: a Deep Learning Method for 21cm Foreground Removal
• Deep learning approach for identification of HII regions during reionization in 21-cm observations
• Constraining the reionization history using deep learning from 21cm tomography with the Square Kilometre Array
• Inferring the properties of the sources of reionization using the morphological spectra of the ionized regions
• HInet: Generating neutral hydrogen from dark matter with neural networks
• Cosmic Velocity Field Reconstruction Using AI
• A unified framework for 21cm tomography sample generation and parameter inference with Progressively Growing GANs
• Measuring the Hubble Constant with cosmic chronometers: a machine learning approach
• Extracting the 21-cm Power Spectrum and the reionization parameters from mock datasets using Artificial Neural Networks
• Improved constraints on reionisation from CMB observations: A parameterisation of the kSZ effect
• Reconstructing Patchy Reionization with Deep Learning
• A New Emulated Monte Carlo Radiative Transfer Disk-Wind Model: X-Ray Accretion Disk-wind Emulator -- XRADE
• Using Artificial Neural Networks to extract the 21-cm Global Signal from the EDGES data
• Deep-Learning Study of the 21cm Differential Brightness Temperature During the Epoch of Reionization
• Neural networks: solving the chemistry of the interstellar medium
• Extracting the Global 21-cm signal from Cosmic Dawn and Epoch of Reionization in the presence of Foreground and Ionosphere
• PINION: Physics-informed neural network for accelerating radiative transfer simulations for cosmic reionization
• A fast method of reionization parameter space exploration using GPR trained SCRIPT
• From few to many maps: A fast map-level emulator for extreme augmentation of CMB systematics datasets
• Reionisation time fields reconstruction from 21 cm signal maps
• Deep Needlet: A CNN based full sky component separation method in Needlet space
• Synthetic Observations with the Square Kilometre Array (SKA) -- development towards an end-to-end pipeline
• Impact of Calibration and Position Errors on Astrophysical Parameters of the HI 21cm Signal
• In Search of Global 21-cm Signal using Artificial Neural Network in light of ARCADE 2
• Cosmological parameter estimation from large-scale structure deep learning
• Predicting Images for the Dynamics Of stellar Clusters (π-DOC): a deep learning framework to predict mass, distance and age of globular clusters
• Accelerating reionization constraints: An ANN-emulator framework for the SCRIPT Semi-numerical Model
• CosmoUiT: A Vision Transformer-UNet Hybrid for Fast and Accurate Emulation of 21-cm Maps from the Epoch of Reionization
• Variational autoencoder for generating realistic $N$-body simulations for dark matter halos