Inferring Cosmological Parameters on SDSS via Domain-Generalized Neural Networks and Lightcone Simulations

Abstract

We present a proof-of-concept simulation-based inference on ${\Omega }_{\mathrm{m}}$ and ${\sigma }_8$ from the SDSS BOSS LOWZ NGC catalog using neural networks and domain generalization techniques without the need of summary statistics. Using rapid lightcone simulations, $\mathrm{L}-\mathrm{PICOLA}$, mock galaxy catalogs are produced that fully incorporate the observational effects. The collection of galaxies is fed as input to a point cloud-based network, $\text{Minkowski-PointNet}$. We also add relatively more accurate $\mathrm{G}\mathrm{ADGET}$ mocks to obtain robust and generalizable neural networks. By explicitly learning the representations which reduces the discrepancies between the two different datasets via the semantic alignment loss term, we show that the latent space configuration aligns into a single plane in which the two cosmological parameters form clear axes. Consequently, during inference, the SDSS BOSS LOWZ NGC catalog maps onto the plane, demonstrating effective generalization and improving prediction accuracy compared to non-generalized models. Results from the ensemble of 25 independently trained machines find ${\Omega }_{\mathrm{m}}=0.339\pm 0.056$ and ${\sigma }_8=0.801\pm 0.061$, inferred only from the distribution of galaxies in the lightcone slices without relying on any indirect summary statistics. A single machine that best adapts to the $\mathrm{G}\mathrm{ADGET}$ mocks yields a tighter prediction of ${\Omega }_{\mathrm{m}}=0.282\pm 0.014$ and ${\sigma }_8=0.786\pm 0.036$. We emphasize that adaptation across multiple domains can enhance the robustness of the neural networks in observational data.

28 pages, 14 figures; Accepted for publication in ApJ

References in corpus (74)

• Stellar population synthesis at the resolution of 2003
• The Sloan Digital Sky Survey: Technical Summary
• Galactic Stellar and Substellar Initial Mass Function
• On the variation of the Initial Mass Function
• The cosmological simulation code GADGET-2
• The 2dF Galaxy Redshift Survey: Spectra and redshifts
• SDSS-III: Massive Spectroscopic Surveys of the Distant Universe, the Milky Way Galaxy, and Extra-Solar Planetary Systems
• The Average Star Formation Histories of Galaxies in Dark Matter Halos from z=0-8
• The Baryon Oscillation Spectroscopic Survey of SDSS-III
• The Cosmic Evolution Survey (COSMOS) -- Overview
• The Rockstar Phase-Space Temporal Halo Finder and the Velocity Offsets of Cluster Cores
• Mass assembly in quiescent and star-forming galaxies since z=4 from UltraVISTA
• Spectroscopic Target Selection for the Sloan Digital Sky Survey: The Luminous Red Galaxy Sample
• KiDS-450: Cosmological parameter constraints from tomographic weak gravitational lensing
• Halo occupation numbers and galaxy bias
• MultiDark simulations: the story of dark matter halo concentrations and density profiles
• The Dark Side of the Halo Occupation Distribution
• The Connection between Galaxies and their Dark Matter Halos
• Modeling Luminosity-Dependent Galaxy Clustering Through Cosmic Time
• KiDS-1000 Cosmology: Cosmic shear constraints and comparison between two point statistics
• Cosmology from cosmic shear power spectra with Subaru Hyper Suprime-Cam first-year data
• Cosmological Parameters from the BOSS Galaxy Power Spectrum
• Simulating cosmic structure formation with the GADGET-4 code
• The Halo Occupation Distribution and the Physics of Galaxy Formation
• The Connection between Galaxies and Dark Matter Structures in the Local Universe
• Solving Large Scale Structure in Ten Easy Steps with COLA
• The Quijote simulations
• The clustering of galaxies in the SDSS-III Baryon Oscillation Spectroscopic Survey: mock galaxy catalogues for the BOSS Final Data Release
• The Clustering of the SDSS Main Galaxy Sample II: Mock galaxy catalogues and a measurement of the growth of structure from Redshift Space Distortions at $z=0.15$
• nbodykit: an open-source, massively parallel toolkit for large-scale structure
• The MICE Grand Challenge Lightcone Simulation I: Dark matter clustering
• Stellar masses of SDSS-III BOSS galaxies at z~0.5 and constraints to galaxy formation models
• The clustering of galaxies in the SDSS-III Baryon Oscillation Spectroscopic Survey: Modeling the clustering and halo occupation distribution of BOSS-CMASS galaxies in the Final Data Release
• The MICE Grand Challenge Lightcone Simulation III: Galaxy lensing mocks from all-sky lensing maps
• Mock galaxy catalogs using the quick particle mesh method
• Methods for Rapidly Processing Angular Masks of Next-Generation Galaxy Surveys
• The MICE Grand Challenge Lightcone Simulation II: Halo and Galaxy catalogues
• The stellar-to-halo mass relation over the past 12 Gyr
• Modelling Baryon Acoustic Oscillations with Perturbation Theory and Stochastic Halo Biasing
• Large-scale Bias and Efficient Generation of Initial Conditions for Non-Local Primordial Non-Gaussianity
• Cosmological constraints with deep learning from KiDS-450 weak lensing maps
• A New Census of the 0.2< z <3.0 Universe, Part II: The Star-Forming Sequence
• A New Census of the 0.2 < z < 3.0 Universe, Part I: The Stellar Mass Function
• Non-Gaussian information from weak lensing data via deep learning
• The clustering of galaxies in the SDSS-III Baryon Oscillation Spectroscopic Survey: galaxy clustering measurements in the low redshift sample of Data Release 11
• A New Method to Correct for Fiber Collisions in Galaxy Two-Point Statistics
• Nuisance hardened data compression for fast likelihood-free inference
• Constraining $M_{\nu }$ with the Bispectrum II: The Total Information Content of the Galaxy Bispectrum
• Modelling galaxy clustering: halo occupation distribution versus subhalo matching
• Cosmological Structure Formation with Augmented Lagrangian Perturbation Theory
• Weak lensing cosmology with convolutional neural networks on noisy data
• Connecting massive galaxies to dark matter halos in BOSS - I. Is galaxy color a stochastic process in high-mass halos?
• Cosmological constraints from noisy convergence maps through deep learning
• The GIGANTES dataset: precision cosmology from voids in the machine learning era
• The Stripe 82 Massive Galaxy Project II: Stellar Mass Completeness of Spectroscopic Galaxy Samples from the Baryon Oscillation Spectroscopic Survey
• Weak Lensing Light-Cones in Modified Gravity simulations with and without Massive Neutrinos
• A Full $w$CDM Analysis of KiDS-1000 Weak Lensing Maps using Deep Learning
• A Hybrid Deep Learning Approach to Cosmological Constraints From Galaxy Redshift Surveys
• Robust Field-level Likelihood-free Inference with Galaxies
• Machine-assisted Semi-Simulation Model (MSSM): Estimating Galactic Baryonic Properties from their Dark Matter using a Machine Trained on Hydrodynamic Simulations
• Cosmology with one galaxy?
• HectoMAP and Horizon Run 4: Dense Structures and Voids in the Real and Simulated Universe
• Cosmological constraints from HSC survey first-year data using deep learning
• $\mathrm{S}\mathrm{IM}\mathrm{BIG}$: Mock Challenge for a Forward Modeling Approach to Galaxy Clustering
• A Subhalo-Galaxy Correspondence Model of Galaxy Formation
• Extracting cosmological parameters from N-body simulations using machine learning techniques
• Inferring Astrophysics and Dark Matter Properties from 21cm Tomography using Deep Learning
• Calibrating cosmological simulations with implicit likelihood inference using galaxy growth observables
• Cosmological Constraints from applying SHAM to rescaled cosmological simulations
• HectoMAP: The Complete Redshift Survey (Data Release 2)
• Robust Neural Network-Enhanced Estimation of Local Primordial Non-Gaussianity
• Constraining cosmological parameters from N-body simulations with Variational Bayesian Neural Networks
• ScamPy -- A sub-halo clustering & abundance matching based Python interface for painting galaxies on the dark matter halo/sub-halo hierarchy
• Artificial Neural Networks for Galaxy Clustering. Learning from the two-point correlation function of BOSS galaxies

Cited by in corpus (3)

• Cosmology with Topological Deep Learning
• 200,000+ Deep Learning-inferred Periods of Stellar Variability from the All-Sky Automated Survey for Supernovae
• Cosmological Analysis with Calibrated Neural Quantile Estimation and Approximate Simulators