Improving constraints on primordial non-Gaussianity using neural network based reconstruction

Abstract

We study the use of U-Nets in reconstructing the linear dark matter density field and its consequences for constraining cosmological parameters, in particular primordial non-Gaussianity. Our network is able to reconstruct the initial conditions of redshift $z=0$ density fields from N-body simulations with $90\%$ accuracy out to $k\le 0.4$ h/Mpc, competitive with state-of-the-art reconstruction algorithms at a fraction of the computational cost. We study the information content of the reconstructed $z=0$ density field with a Fisher analysis using the QUIJOTE simulation suite, including non-Gaussian initial conditions. Combining the pre- and post-reconstructed power spectrum and bispectrum data up to $k_{\max }=0.52$ h/Mpc, we find significant improvements on all parameters. Most notably, we find a factor $3.65$ (local), $3.54$ (equilateral) and $2.90$ (orthogonal) improvement on the marginalized errors of $f_{\mathrm{NL}}$ as compared to only using the pre-reconstructed data. We show that these improvements can be attributed to a combination of reduced data covariance and parameter degeneracy. The results constitute an important step towards more optimal inference of primordial non-Gaussianity from non-linear scales.

arXiv comment22 pages, 8 figures, 3 tables, codes available at https://github.com/tsfloss/URecon and https://github.com/tsfloss/DensityFieldTools. v2 matches version accepted for JCAP

References in corpus (59)

• Non-Gaussian features of primordial fluctuations in single field inflationary models
• Large-Scale Structure of the Universe and Cosmological Perturbation Theory
• The Simons Observatory: Science goals and forecasts
• The Effective Field Theory of Inflation
• 21-cm cosmology
• Why your model parameter confidences might be too optimistic -- unbiased estimation of the inverse covariance matrix
• First Year Wilkinson Microwave Anisotropy Probe (WMAP) Observations: Tests of Gaussianity
• Single field consistency relation for the 3-point function
• The Effective Field Theory of Cosmological Large Scale Structures
• Improving Cosmological Distance Measurements by Reconstruction of the Baryon Acoustic Peak
• Planck 2015 results. XVII. Constraints on primordial non-Gaussianity
• The shape of non-Gaussianities
• Constraints on local primordial non-Gaussianity from large scale structure
• Non-Gaussianities in Single Field Inflation and their Optimal Limits from the WMAP 5-year Data
• Limits on non-Gaussianities from WMAP data
• The Quijote simulations
• Correcting for the alias effect when measuring the power spectrum using FFT
• Cosmological Collider Physics
• Non-Gaussianity as a Particle Detector
• Accurate Estimators of Correlation Functions in Fourier Space
• Learning to Predict the Cosmological Structure Formation
• RegPT: Direct and fast calculation of regularized cosmological power spectrum at two-loop order
• The Bispectrum: From Theory to Observations
• The Effective Field Theory of Large Scale Structures at Two Loops
• Line-Intensity Mapping: 2017 Status Report
• The Clustering of Galaxies in SDSS-III DR9 Baryon Oscillation Spectroscopic Survey: Constraints on Primordial Non-Gaussianity
• Evidence for non-Gaussianity in the COBE DMR Four Year Sky Maps
• Measurement of the cosmic microwave background bispectrum on the COBE DMR sky maps
• Redshift-weighted constraints on primordial non-Gaussianity from the clustering of the eBOSS DR14 quasars in Fourier space
• Prospects for Cosmological Collider Physics
• Constraints on primordial non-Gaussianity from 800,000 photometric quasars
• Towards an Optimal Estimation of Cosmological Parameters with the Wavelet Scattering Transform
• Primordial non-gaussianity from the bispectrum of 21-cm fluctuations in the dark ages
• The One-Loop Matter Bispectrum in the Effective Field Theory of Large Scale Structures
• Iterative initial condition reconstruction
• Assessment of the Information Content of the Power Spectrum and Bispectrum
• Eulerian BAO Reconstructions and N-Point Statistics
• Inflation: Theory and Observations
• deep21: a Deep Learning Method for 21cm Foreground Removal
• Sloan Digital Sky Survey III Photometric Quasar Clustering: Probing the Initial Conditions of the Universe using the Largest Volume
• GridSPT: Grid-based calculation for perturbation theory of large-scale structure
• Recovering the Wedge Modes Lost to 21-cm Foregrounds
• Field Level Neural Network Emulator for Cosmological N-body Simulations
• Quijote-PNG: Simulations of primordial non-Gaussianity and the information content of the matter field power spectrum and bispectrum
• Constraining Single-Field Inflation with MegaMapper
• The Covariance of Squeezed Bispectrum Configurations
• Reconstructing Cosmological Initial Conditions from Late-Time Structure with Convolutional Neural Networks
• Constraining Primordial Non-Gaussianity with Post-reconstructed Galaxy Bispectrum in Redshift Space
• Fundamental limits on constraining primordial non-Gaussianity
• Minimizing gravitational lensing contributions to the primordial bispectrum covariance
• Quijote-PNG: Quasi-maximum likelihood estimation of Primordial Non-Gaussianity in the non-linear dark matter density field
• Effective cosmic density field reconstruction with convolutional neural network
• Probing primordial non-Gaussianity with the power spectrum and bispectrum of future 21cm intensity maps
• Primordial non-Gaussianity and non-Gaussian Covariance
• The Dark Ages' 21-cm Trispectrum
• Squeezing $f_{\mathrm{NL}}$ out of the matter bispectrum with consistency relations
• Simple lessons from complex learning: what a neural network model learns about cosmic structure formation
• A Model for the Squeezed Bispectrum in the Non-Linear Regime
• Predicting the Initial Conditions of the Universe using a Deterministic Neural Network

Cited by in corpus (8)

• Capturing primordial non-Gaussian signatures in the late Universe by multi-scale extrema of the cosmic log-density field
• Deep learning approach for identification of HII regions during reionization in 21-cm observations -- III. image recovery
• Learning the Universe: Learning to Optimize Cosmic Initial Conditions with Non-Differentiable Structure Formation Models
• Neural Network Reconstruction of Non-Gaussian Initial Conditions from Dark Matter Halos
• Modelling the covariance matrix for the power spectra before and after the BAO reconstruction
• Power Spectrum Emulators from Neural Networks and Tree-Based Methods
• Towards detecting Primordial non-Gaussianity in the CMB using Spherical Convolutional Neural Networks
• Fast Sampling of Cosmological Initial Conditions with Gaussian Neural Posterior Estimation