Reducing nuisance prior sensitivity via non-linear reparameterization, with application to EFT analyses of large-scale structure

Abstract

Many physical models contain nuisance parameters that quantify unknown properties of an experiment that are not of primary relevance. Typically, these cannot be measured except by fitting the models to the data from the experiment, requiring simultaneous measurement of interesting parameters that are our target of inference and nuisance terms that are not directly of interest. A recent example of this is fitting Effective Field Theory (EFT) models to large-scale structure (LSS) data to make cosmological inferences. These models have a large number of nuisance parameters that are typically correlated with cosmological parameters in the posterior, leading to strong dependence on the nuisance parameter priors. We introduce a reparametrization method that leverages Generalized Additive Models (GAMs) to decorrelate nuisance parameters from the parameters of interest in the likelihood, even in the presence of non-linear relationships. This reparametrization forms a natural basis within which to define priors that are independent between nuisance and target parameters: the separation means that the marginal posterior for cosmological parameters does not depend on simple priors placed on nuisance terms. In application to EFT models using LSS data, we demonstrate that the proposed approach leads to robust cosmological inference.

27 pages, 11 figures, 5 tables

References in corpus (26)

• The DESI Experiment Part I: Science,Targeting, and Survey Design
• Bayes in the sky: Bayesian inference and model selection in cosmology
• The Effective Field Theory of Cosmological Large Scale Structures
• Overview of the Instrumentation for the Dark Energy Spectroscopic Instrument
• A General Framework for the Parametrization of Hierarchical Models
• The IR-resummed Effective Field Theory of Large Scale Structures
• The DESI Experiment Part II: Instrument Design
• The BOSS bispectrum analysis at one loop from the Effective Field Theory of Large-Scale Structure
• Redshift Space Distortions in the Effective Field Theory of Large Scale Structures
• Cosmology From CMB Lensing and Delensed EE Power Spectra Using 2019-2020 SPT-3G Polarization Data
• Consistency of effective field theory analyses of the BOSS power spectrum
• Accelerating astronomical and cosmological inference with Preconditioned Monte Carlo
• Partially Acoustic Dark Matter Cosmology and Cosmological Constraints
• JAX-COSMO: An End-to-End Differentiable and GPU Accelerated Cosmology Library
• Cosmological inference from the EFTofLSS: the eBOSS QSO full-shape analysis
• Cosmological constraints from the power spectrum of eBOSS quasars
• Cosmology with 6 parameters in the Stage-IV era: efficient marginalisation over nuisance parameters
• CosmoPower-JAX: high-dimensional Bayesian inference with differentiable cosmological emulators
• Capse.jl: efficient and auto-differentiable CMB power spectra emulation
• candl: Cosmic Microwave Background Analysis with a Differentiable Likelihood
• EFTofLSS meets simulation-based inference: ${\sigma }_8$ from biased tracers
• Analysis of Unified Galaxy Power Spectrum Multipole Measurements
• LimberJack.jl: auto-differentiable methods for angular power spectra analyses
• Effort: a fast and differentiable emulator for the Effective Field Theory of the Large Scale Structure of the Universe
• On priors and scale cuts in EFT-based full-shape analyses
• Generalized Multivariate Functional Additive Mixed Models for Location, Scale, and Shape