Bayesian deep learning for cosmic volumes with modified gravity

Abstract

The new generation of galaxy surveys will provide unprecedented data allowing us to test gravity at cosmological scales. A robust cosmological analysis of the large-scale structure demands exploiting the nonlinear information encoded in the cosmic web. Machine Learning techniques provide such tools, however, do not provide a priori assessment of uncertainties. This study aims at extracting cosmological parameters from modified gravity (MG) simulations through deep neural networks endowed with uncertainty estimations. We implement Bayesian neural networks (BNNs) with an enriched approximate posterior distribution considering two cases: one with a single Bayesian last layer (BLL), and another one with Bayesian layers at all levels (FullB). We train both BNNs with real-space density fields and power-spectra from a suite of 2000 dark matter only particle mesh $N$-body simulations including modified gravity models relying on MG-PICOLA covering 256 $h^{-1}$ Mpc side cubical volumes with 128${}^3$ particles. BNNs excel in accurately predicting parameters for ${\Omega }_m$ and ${\sigma }_8$ and their respective correlation with the MG parameter. We find out that BNNs yield well-calibrated uncertainty estimates overcoming the over- and under-estimation issues in traditional neural networks. We observe that the presence of MG parameter leads to a significant degeneracy with ${\sigma }_8$ being one of the possible explanations of the poor MG predictions. Ignoring MG, we obtain a deviation of the relative errors in ${\Omega }_m$ and ${\sigma }_8$ by at least $30\%$. Moreover, we report consistent results from the density field and power spectra analysis, and comparable results between BLL and FullB experiments which permits us to save computing time by a factor of two. This work contributes in setting the path to extract cosmological parameters from complete small cosmic volumes towards the highly nonlinear regime.

13 pages, 7 figures and 7 tables. Minor changes to match the accepted version in A&A

References in corpus (67)

• f(R) theories
• Modified Gravity Theories on a Nutshell: Inflation, Bounce and Late-time Evolution
• A Review of Uncertainty Quantification in Deep Learning: Techniques, Applications and Challenges
• Models of f(R) Cosmic Acceleration that Evade Solar-System Tests
• Testing General Relativity with Present and Future Astrophysical Observations
• Observational Probes of Cosmic Acceleration
• Transients from Initial Conditions in Cosmological Simulations
• Cosmology with cosmic shear observations: a review
• The Large Scale Structure of f(R) Gravity
• Observational signatures of f(R) dark energy models that satisfy cosmological and local gravity constraints
• Solving Large Scale Structure in Ten Easy Steps with COLA
• The Quijote simulations
• ECOSMOG: An Efficient Code for Simulating Modified Gravity
• Recent Advances on Inflation
• The Three-Point Correlation Function in Cosmology
• Cosmic Shear Statistics and Cosmology
• Constraints on Cosmology and Gravity from the Dynamics of Voids
• Testing Gravity using Cosmic Voids
• Modified Gravity-GADGET: A new code for cosmological hydrodynamical simulations of modified gravity models
• The darkness that shaped the void: dark energy and cosmic voids
• Constraints on scalar-tensor models of dark energy from observational and local gravity tests
• Cosmology constraints from shear peak statistics in Dark Energy Survey Science Verification data
• Precision cosmology with voids: definition, methods, dynamics
• Redshift space distortions in f(R) gravity
• Probing Cosmology with Weak Lensing Minkowski Functionals
• Minkowski Functionals of SDSS galaxies I : Analysis of Excursion Sets
• COLA with scale-dependent growth: applications to screened modified gravity models
• Evidence for suppression of structure growth in the concordance cosmological model
• Cosmic Shear Cosmology Beyond 2-Point Statistics: A Combined Peak Count and Correlation Function Analysis of DES-Y1
• Fast and accurate mock catalogue generation for low-mass galaxies
• Cosmic voids in modified gravity models with massive neutrinos
• Nonlinear growth in modified gravity theories of dark energy
• Super-resolution emulator of cosmological simulations using deep physical models
• Distinguishing standard and modified gravity cosmologies with machine learning
• Modelling Void Abundance in Modified Gravity
• Redshift-space distortions with split densities
• Large-Scale Structure of the Universe as a Cosmic Standard Ruler
• Cosmic voids in modified gravity scenarios
• Parameters Estimation for the Cosmic Microwave Background with Bayesian Neural Networks
• Searching for Modified Gravity: Scale and Redshift Dependent Constraints from Galaxy Peculiar Velocities
• Cosmological constraints with weak lensing peak counts and second-order statistics in a large-field survey
• COSMIC BIRTH: Efficient Bayesian Inference of the Evolving Cosmic Web from Galaxy Surveys
• ENCORE: An $\mathcal{O}(N_{\mathrm{g}}^2)$ Estimator for Galaxy $N$-Point Correlation Functions
• Joint halo mass function for modified gravity and massive neutrinos I: simulations and cosmological forecasts
• Large-scale redshift space distortions in modified gravity theories
• On the dissection of degenerate cosmologies with machine learning
• Consistency tests of field level inference with the EFT likelihood
• New probe of departures from general relativity using Minkowski functionals
• Fast full $N$-body simulations of generic modified gravity: derivative coupling models
• Tomographic Alcock-Paczynski Test with Redshift-Space Correlation Function: Evidence for the Dark Energy Equation of State Parameter w>-1
• Non-parametric dark energy reconstruction using the tomographic Alcock-Paczynski test
• Cosmological constraints from the redshift dependence of the Alcock-Paczynski test and volume effect: galaxy two-point correlation function
• Weak lensing peak statistics in the era of large scale cosmological surveys
• Variational Inference as an alternative to MCMC for parameter estimation and model selection
• Clustering and redshift-space distortions in modified gravity models with massive neutrinos
• Extracting cosmological parameters from N-body simulations using machine learning techniques
• Improved uncertainty quantification for neural networks with Bayesian last layer
• Investigating Cosmological GAN Emulators Using Latent Space Interpolation
• Probing gravity with redshift-space distortions: effects of tracer bias and sample selection
• Distinguishing screening mechanisms with environment-dependent velocity statistics
• Validating the methodology for constraining the linear growth rate from clustering anisotropies
• Constraining cosmological parameters from N-body simulations with Variational Bayesian Neural Networks
• The halo 3-point correlation function: a methodological analysis
• Constraining galaxy-halo connection with high-order statistics
• Testing modified gravity scenarios with direct peculiar velocities
• ConKer: evaluating isotropic correlations of arbitrary order
• Cosmological constraints from the redshift dependence of the Alcock-Paczynski effect: Fourier space analysis

Cited by in corpus (1)

• Neural Networks for cosmological model selection and feature importance using Cosmic Microwave Background data