اشترك بالحزمة الذهبية واحصل على وصول غير محدود شمرا أكاديميا
تسجيل مستخدم جديدCosmological Parameter Estimation and Inference using Deep Summaries
534
0
0.0
(
0
)
اسأل ChatGPT حول البحث
ﻻ يوجد ملخص باللغة العربية
The ability to obtain reliable point estimates of model parameters is of crucial importance in many fields of physics. This is often a difficult task given that the observed data can have a very high number of dimensions. In order to address this problem, we propose a novel approach to construct parameter estimators with a quantifiable bias using an order expansion of highly compressed deep summary statistics of the observed data. These summary statistics are learned automatically using an information maximising loss. Given an observation, we further show how one can use the constructed estimators to obtain approximate Bayes computation (ABC) posterior estimates and their corresponding uncertainties that can be used for parameter inference using Gaussian process regression even if the likelihood is not tractable. We validate our method with an application to the problem of cosmological parameter inference of weak lensing mass maps. We show in that case that the constructed estimators are unbiased and have an almost optimal variance, while the posterior distribution obtained with the Gaussian process regression is close to the true posterior and performs better or equally well than comparable methods.
قيم البحث
اقرأ أيضاً
Cosmological parameter estimation is entering a new era. Large collaborations need to coordinate high-stakes analyses using multiple methods; furthermore such analyses have grown in complexity due to sophisticated models of cosmology and systematic u
ncertainties. In this paper we argue that modularity is the key to addressing these challenges: calculations should be broken up into interchangeable modular units with inputs and outputs clearly defined. We present a new framework for cosmological parameter estimation, CosmoSIS, designed to connect together, share, and advance development of inference tools across the community. We describe the modules already available in CosmoSIS, including CAMB, Planck, cosmic shear calculations, and a suite of samplers. We illustrate it using demonstration code that you can run out-of-the-box with the installer available at http://bitbucket.org/joezuntz/cosmosis
We study the topology of the matter density field in two dimensional slices, and consider how we can use the amplitude $A$ of the genus for cosmological parameter estimation. Using the latest Horizon Run 4 simulation data, we calculate the genus of t
he smoothed density field constructed from lightcone mock galaxy catalogs. Information can be extracted from the amplitude of the genus by considering both its redshift evolution and magnitude. The constancy of the genus amplitude with redshift can be used as a standard population, from which we derive constraints on the equation of state of dark energy $w_{rm de}$ - by measuring $A$ at $z sim 0.1$ and $z sim 1$, we can place an order $Delta w_{rm de} sim {cal O}(15%)$ constraint on $w_{rm de}$. By comparing $A$ to its Gaussian expectation value we can potentially derive an additional stringent constraint on the matter density $Delta Omega_{rm mat} sim 0.01$. We discuss the primary sources of contamination associated with the two measurements - redshift space distortion and shot noise. With accurate knowledge of galaxy bias, we can successfully remove the effect of redshift space distortion, and the combined effect of shot noise and non-linear gravitational evolution is suppressed by smoothing over suitably large scales $R_{rm G} ge 15 {rm Mpc}/h$. Without knowledge of the bias, we discuss how joint measurements of the two and three dimensional genus can be used to constrain the growth factor $beta = f/b$. The method can be applied optimally to redshift slices of a galaxy distribution generated using the drop-off technique.
We use mock galaxy survey simulations designed to resemble the Dark Energy Survey Year 1 (DES Y1) data to validate and inform cosmological parameter estimation. When similar analysis tools are applied to both simulations and real survey data, they pr
ovide powerful validation tests of the DES Y1 cosmological analyses presented in companion papers. We use two suites of galaxy simulations produced using different methods, which therefore provide independent tests of our cosmological parameter inference. The cosmological analysis we aim to validate is presented in DES Collaboration et al. (2017) and uses angular two-point correlation functions of galaxy number counts and weak lensing shear, as well as their cross-correlation, in multiple redshift bins. While our constraints depend on the specific set of simulated realisations available, for both suites of simulations we find that the input cosmology is consistent with the combined constraints from multiple simulated DES Y1 realizations in the $Omega_m-sigma_8$ plane. For one of the suites, we are able to show with high confidence that any biases in the inferred $S_8=sigma_8(Omega_m/0.3)^{0.5}$ and $Omega_m$ are smaller than the DES Y1 $1-sigma$ uncertainties. For the other suite, for which we have fewer realizations, we are unable to be this conclusive; we infer a roughly 70% probability that systematic biases in the recovered $Omega_m$ and $S_8$ are sub-dominant to the DES Y1 uncertainty. As cosmological analyses of this kind become increasingly more precise, validation of parameter inference using survey simulations will be essential to demonstrate robustness.
We assess the performance of a perturbation theory inspired method for inferring cosmological parameters from the joint measurements of galaxy-galaxy weak lensing ($DeltaSigma$) and the projected galaxy clustering ($w_{rm p}$). To do this, we use a w
ide variety of mock galaxy catalogs constructed based on a large set of $N$-body simulations that mimic the Subaru HSC-Y1 and SDSS galaxies, and apply the method to the mock signals to address whether to recover the underlying true cosmological parameters in the mocks. We find that, as long as the appropriate scale cuts, $12$ and $8~h^{-1}{rm Mpc}$ for $DeltaSigma$ and $w_{rm p}$ respectively, are adopted, a minimal-bias model using the linear bias parameter $b_1$ alone and the nonlinear matter power spectrum can recover the true cosmological parameters (here focused on $Omega_{rm m}$ and $sigma_8$) to within the 68% credible interval, for all the mocks we study including one in which an assembly bias effect is implemented. This is as expected if physical processes inherent in galaxy formation/evolution are confined to local, small scales below the scale cut, and thus implies that real-space observables have an advantage in filtering out the impact of small-scale nonlinear effects in parameter estimation, compared to their Fourier-space counterparts. In addition, we find that a theoretical template including the higher-order bias contributions such as nonlinear bias parameter $(b_2)$ does not improve the cosmological constraints, but rather leads to a larger parameter bias compared to the baseline $b_1$-method.
Cosmological large-scale structure analyses based on two-point correlation functions often assume a Gaussian likelihood function with a fixed covariance matrix. We study the impact on cosmological parameter estimation of ignoring the parameter depend
ence of this covariance matrix, focusing on the particular case of joint weak-lensing and galaxy clustering analyses. Using a Fisher matrix formalism (calibrated against exact likelihood evaluation in particular simple cases), we quantify the effect of using a parameter dependent covariance matrix on both the bias and variance of the parameters. We confirm that the approximation of a parameter-independent covariance matrix is exceptionally good in all realistic scenarios. The information content in the covariance matrix (in comparison with the two point functions themselves) does not change with the fractional sky coverage. Therefore the increase in information due to the parameter dependent covariance matrix becomes negligible as the number of modes increases. Even for surveys covering less than $1%$ of the sky, this effect only causes a bias of up to ${cal O}(10%)$ of the statistical uncertainties, with a misestimation of the parameter uncertainties at the same level or lower. The effect will only be smaller with future large-area surveys. Thus for most analyses the effect of a parameter-dependent covariance matrix can be ignored both in terms of the accuracy and precision of the recovered cosmological constraints.
سجل دخول لتتمكن من نشر تعليقات
التعليقات
جاري جلب التعليقات
سجل دخول لتتمكن من متابعة معايير البحث التي قمت باختيارها
