$\texttt{matryoshka}$ II: Accelerating Effective Field Theory Analyses of the Galaxy Power Spectrum

Abstract

In this paper we present an extension to the matryoshka suite of neural network based emulators. The new editions have been developed to accelerate EFTofLSS analyses of galaxy power spectrum multipoles in redshift space. They are collectively referred to as the EFTEMU. We test the EFTEMU at the power spectrum level and achieve a prediction accuracy of better than 1% with BOSS-like bias parameters and counterterms on scales 0.001 ℎ Mpc −1 𝑘 0.19 ℎ Mpc −1 . We also run a series of mock full shape analyses to test the EFTEMU at the inference level. Through these mock analyses we verify that the EFTEMU recovers the true cosmology within 1𝜎 at several redshifts (𝑧 = [0.38, 0.51, 0.61]), and with several noise levels (the most stringent of which being a Gaussian covariance associated with a volume of 5000 3 Mpc 3 ℎ −3 ). We compare mock inference results from the EFTEMU to those obtained with a fully analytic EFTofLSS model and again find no significant bias, whilst speeding up the inference by three orders of magnitude. The EFTEMU is publicly available as part of the matryoshka Python package https://github.com/JDonaldM/Matryoshka.

Machine Learning for Observational Cosmology

“…We list large-scale cosmological simulation projects for emulators that interpolate summary statistics measured from simulations over the cosmological parameter space in Table 1. We note also that there are many other attempts to utilize emulators for different purposes, such as developing fast Boltzmann equation solvers or performing low-order perturbative calculations [323][324][325][326][327][328][329][330][331][332][333][334][335][336], to explore the galaxy-halo connection for fixed cosmology [337], to translate less costly, low-resolution simulations to mimic more expensive simulations. More specifically, the applications include incorporating baryonic effects to the dark-matter only simulations [338], predicting Lyman-α forest [339,340] or 21-cm power spectra [341,342], extrapolating the predictions in Λ-Cold Dark Matter ¶ Here, convergence means the change in the observed size of an object caused by gravitational lensing effect.…”

COMET: Clustering Observables Modelled by Emulated perturbation Theory

“…While most galaxy clustering emulators that have been presented in the literature so far have been built from simulations and focus on the non-linear regime, there are two sets of works, which are closely related to what we have presented here and which we want to briefly compare against. In particular, Donald-McCann et al (2022); DeRose et al ( 2022) have presented two perturbation theory emulators of the galaxy power spectrum multipoles. The former, EFTEMU, is based on the PyBird code (d' Amico et al 2020), which implements perturbation theory expressions identical to those described in Sec.…”

“…Chudaykin et al 2020;Chen et al 2021), this still amounts to considerable computational costs. We intend to greatly facilitate this task by constructing an emulator for various clustering statistics in perturbation theory, in particular the multipoles of the redshift-space galaxy power spectrum, that is capable of drastically reducing the time needed for making model predictions (see DeRose et al 2022;Donald-McCann et al 2022, for a similar approach). We thus follow a recent string of works, which introduced emulators as a means of overcoming the much higher computational costs of running simulations, which are needed to extend models of galaxy clustering into the deeply non-linear regime.…”

“…In order to satisfy that demand it is important to keep the parameter space over which the emulator is build as small as possible. As in Donald-McCann et al (2022); Aricò et al (2021), we make use of the fact that all model parameters related to the galaxy bias expansion can be factorised and so their dependency does not need to be emulated. However, what sets our emulator apart from these other works is that it is build around the recently proposed evolution mapping approach (Sánchez et al 2022), which further limits the emulation parameter space.…”

Updated neutrino mass constraints from galaxy clustering and CMB lensing-galaxy cross-correlation measurements

“…5 Other works have applied related perturbation theory-based models (see e.g. [142,143]) for the redshift-space galaxy power spectrum to real data [144,145] or forecasts [146][147][148][149][150][151], tested these models on simulations [152][153][154][155], or investigated other ways of extracting (possibly compressed) information from the redshift-space FS galaxy power spectrum and higher order multipoles [156][157][158][159][160][161][162][163][164][165][166][167]. Finally, various works have used simulations to investigate the imprint of neutrino properties on the clustering of the LSS (see for e.g.…”