Robust Neural Network-Enhanced Estimation of Local Primordial Non-Gaussianity

Abstract

When applied to the non-linear matter distribution of the universe, neural networks have been shown to be very statistically sensitive probes of cosmological parameters, such as the linear perturbation amplitude ${\sigma }_8$. However, when used as a "black box", neural networks are not robust to baryonic uncertainty. We propose a robust architecture for constraining primordial non-Gaussianity $f_{NL}$, by training a neural network to locally estimate ${\sigma }_8$, and correlating these local estimates with the large-scale density field. We apply our method to N-body simulations, and show that $\sigma (f_{NL})$ is 3.5 times better than the constraint obtained from a standard halo-based approach. We show that our method has the same robustness property as large-scale halo bias: baryonic physics can change the normalization of the estimated $f_{NL}$, but cannot change whether $f_{NL}$ is detected.

References in corpus (28)

• Array Programming with NumPy
• emcee: The MCMC Hammer
• The Cosmic Linear Anisotropy Solving System (CLASS) II: Approximation schemes
• Non-Gaussianity from Inflation: Theory and Observations
• The imprints of primordial non-gaussianities on large-scale structure: scale dependent bias and abundance of virialized objects
• Large-Scale Galaxy Bias
• Nongaussian Isocurvature Perturbations from Inflation
• A new mechanism for generating density perturbations from inflation
• Constraints on local primordial non-Gaussianity from large scale structure
• Measuring primordial non-gaussianity without cosmic variance
• The Quijote simulations
• Bayesian physical reconstruction of initial conditions from large scale structure surveys
• The CAMELS project: Cosmology and Astrophysics with MachinE Learning Simulations
• Galaxy Bias and non-Linear Structure Formation in General Relativity
• Towards an Optimal Estimation of Cosmological Parameters with the Wavelet Scattering Transform
• Towards optimal extraction of cosmological information from nonlinear data
• Constraining local non-Gaussianities with kSZ tomography
• Primordial Non-Gaussianity from Biased Tracers: Likelihood Analysis of Real-Space Power Spectrum and Bispectrum
• Verifying the consistency relation for the scale-dependent bias from local primordial non-Gaussianity
• Scale-dependent bias and bispectrum in neutrino separate universe simulations
• Going Beyond the Galaxy Power Spectrum: an Analysis of BOSS Data with Wavelet Scattering Transforms
• Learning cosmology and clustering with cosmic graphs
• The Hunt for Primordial Interactions in the Large Scale Structures of the Universe
• Halo clustering and g_{NL}-type primordial non-Gaussianity
• A Hybrid Deep Learning Approach to Cosmological Constraints From Galaxy Redshift Surveys
• Baryon-CDM isocurvature galaxy bias with IllustrisTNG
• Exploring KSZ velocity reconstruction with $N$-body simulations and the halo model
• Extracting cosmological parameters from N-body simulations using machine learning techniques

Cited by in corpus (5)

• Learning to Concentrate: Multi-tracer Forecasts on Local Primordial Non-Gaussianity with Machine-Learned Bias
• Quijote-PNG: Quasi-maximum likelihood estimation of Primordial Non-Gaussianity in the non-linear halo density field
• Capturing primordial non-Gaussian signatures in the late Universe by multi-scale extrema of the cosmic log-density field
• Inferring Cosmological Parameters on SDSS via Domain-Generalized Neural Networks and Lightcone Simulations
• Constraining $f_{NL}$ using the Large-Scale Modulation of Small-Scale Statistics