Imperial College London

ProfessorAlanHeavens

Faculty of Natural SciencesDepartment of Physics

Chair in Astrostatistics
 
 
 
//

Contact

 

+44 (0)20 7594 2930a.heavens Website

 
 
//

Location

 

1018EBlackett LaboratorySouth Kensington Campus

//

Summary

 

Publications

Citation

BibTex format

@article{Leclercq:2019:mnras/stz2718,
author = {Leclercq, F and Enzi, W and Jasche, J and Heavens, A},
doi = {mnras/stz2718},
journal = {Monthly Notices of the Royal Astronomical Society},
pages = {4237--4253},
title = {Primordial power spectrum and cosmology from black-box galaxy surveys},
url = {http://dx.doi.org/10.1093/mnras/stz2718},
volume = {490},
year = {2019}
}

RIS format (EndNote, RefMan)

TY  - JOUR
AB - We propose a new, likelihood-free approach to inferring the primordial matterpower spectrum and cosmological parameters from arbitrarily complex forwardmodels of galaxy surveys where all relevant statistics can be determined fromnumerical simulations, i.e. black-boxes. Our approach builds upon approximateBayesian computation using a novel effective likelihood, and upon thelinearisation of black-box models around an expansion point. Consequently, weobtain simple "filter equations" for an effective posterior of the primordialpower spectrum, and a straightforward scheme for cosmological parameterinference. We demonstrate that the workload is computationally tractable, fixeda priori, and perfectly parallel. As a proof of concept, we apply our frameworkto a realistic synthetic galaxy survey, with a data model accounting forphysical structure formation and incomplete and noisy galaxy observations. Indoing so, we show that the use of non-linear numerical models allows the galaxypower spectrum to be safely fitted up to at least $k_\mathrm{max} = 0.5$$h$/Mpc, outperforming state-of-the-art backward-modelling techniques by afactor of $\sim 5$ in the number of modes used. The result is an unbiasedinference of the primordial matter power spectrum across the entire range ofscales considered, including a high-fidelity reconstruction of baryon acousticoscillations. It translates into an unbiased and robust inference ofcosmological parameters. Our results pave the path towards easy applications oflikelihood-free simulation-based inference in cosmology.
AU - Leclercq,F
AU - Enzi,W
AU - Jasche,J
AU - Heavens,A
DO - mnras/stz2718
EP - 4253
PY - 2019///
SN - 0035-8711
SP - 4237
TI - Primordial power spectrum and cosmology from black-box galaxy surveys
T2 - Monthly Notices of the Royal Astronomical Society
UR - http://dx.doi.org/10.1093/mnras/stz2718
UR - https://academic.oup.com/mnras/advance-article/doi/10.1093/mnras/stz2718/5583010
UR - http://hdl.handle.net/10044/1/73943
VL - 490
ER -