arXiv subjects

Guilhem Lavaux

Publications and source records attributed to Guilhem Lavaux.

At least 73 records · Page 4Linked to original sources

arXiv2020

Setting the Stage: Structures from Gaussian Random Fields

We study structure formation in a set of cosmological simulations to uncover the scales in the initial density field that gave rise to the formation of present-day structures. Our simulations share a common primordial power spectrum (here Lambda-CDM), but the introduction of hierarchical variations of the phase information allows us to systematically study the scales that determine the formation of structure at later times. We consider the variance in z=0 statistics such as the matter power spectrum and halo mass function. We also define a criterion for the existence of individual haloes across simulations, and determine what scales in the initial density field contain sufficient information for the non-linear formation of unique haloes. We study how the characteristics of individual haloes such as the mass and concentration, as well as the position and velocity, are affected by variations on different scales, and give scaling relations for haloes of different mass. Finally, we use the example of a cluster-mass halo to show how our hierarchical parametrisation of the initial density field can be used to create variants of particular objects. With properties such as mass, concentration, kinematics and substructure of haloes set on distinct and well-determined scales, and its unique ability to introduce variations localised in real space, our method is a powerful tool to study structure formation in cosmological simulations.

astro-ph.CO↗

arXiv2020

Cosmology Inference from Biased Tracers using the EFT-based Likelihood

The effective-field-theory (EFT) approach to the clustering of galaxies and other biased tracers allows for an isolation of the cosmological information that is protected by symmetries, in particular the equivalence principle, and thus is robust to the complicated dynamics of dark matter, gas, and stars on small scales. All existing implementations proceed by making predictions for the lowest-order $n$-point functions of biased tracers, as well as their covariance, and comparing with measurements. Recently, we presented an EFT-based expression for the conditional probability of the density field of a biased tracer given the matter density field, which in principle combines information from arbitrarily high order $n$-point functions. Here, we report results based on this likelihood by applying it to halo catalogs in real space, specifically on the inference of the power spectrum normalization $σ_8$. We include bias terms up to second order as well as the leading higher-derivative term. For a cutoff value of $Λ= 0.1 h\,{\rm Mpc}^{-1}$, we recover the ground-truth value of $σ_8$ to within 95% CL for different halo samples and redshifts. We discuss possible sources for the remaining systematic bias in $σ_8$ as well as future developments.

astro-ph.CO↗

arXiv2019

The primordial magnetic field in our cosmic backyard

We reconstruct the 3D structure of magnetic fields, which were seeded by density perturbations during the radiation dominated epoch of the Universe and later on were evolved by structure formation. To achieve this goal, we rely on three dimensional initial density fields inferred from the 2M++ galaxy compilation via the Bayesian $\texttt{BORG}$ algorithm. Using those, we estimate the magnetogenesis by the so called Harrison mechanism. This effect produced magnetic fields exploiting the different photon drag on electrons and ions in vortical motions, which are exited due to second order perturbation effects in the Early Universe. Subsequently we study the evolution of these seed fields through the non-linear cosmic structure formation by virtue of a MHD simulation to obtain a 3D estimate for the structure of this primordial magnetic field component today. At recombination we obtain a reliable lower limit on the large scale magnetic field strength around $10^{-23} \mathrm{G}$, with a power spectrum peaking at about $ 2\, \mathrm{Mpc}^{-1}h$ in comoving scales. At present we expect this evolved primordial field to have strengthts above $\approx 10^{-27}\, \mathrm{G}$ and $\approx 10^{-29}\, \mathrm{G}$ in clusters of galaxies and voids, respectively. We also calculate the corresponding Faraday rotation measure map and show the magnetic field morphology and strength for specific objects of the Local Universe.

astro-ph.CO↗

arXiv2019

Wiener filtering and pure E/B decomposition of CMB maps with anisotropic correlated noise

We present an augmented version of our dual messenger algorithm for spin field reconstruction on the sphere, while accounting for highly non-trivial and realistic noise models such as modulated correlated noise. We also describe an optimization method for the estimation of noise covariance from Monte Carlo simulations. Using simulated Planck polarized cosmic microwave background (CMB) maps as a showcase, we demonstrate the capabilities of the algorithm in reconstructing pure E and B maps, guaranteed to be free from ambiguous modes resulting from the leakage or coupling issue that plagues conventional methods of E/B separation. Due to its high speed execution, coupled with lenient memory requirements, the algorithm can be optimized in exact global Bayesian analyses of state-of-the-art CMB data for a statistically optimal separation of pure E and B modes. Our algorithm, therefore, has a potentially key role in the data analysis of high-resolution and high-sensitivity CMB data, especially with the range of upcoming CMB experiments tailored for the detection of the elusive primordial B-mode signal.

astro-ph.CO↗

arXiv2019

Inferring high redshift large-scale structure dynamics from the Lyman-alpha forest

One of the major science goals over the coming decade is to test fundamental physics with probes of the cosmic large-scale structure out to high redshift. Here we present a fully Bayesian approach to infer the three-dimensional cosmic matter distribution and its dynamics at $z>2$ from observations of the Lyman-$α$ forest. We demonstrate that the method recovers the unbiased mass distribution and the correct matter power spectrum at all scales. Our method infers the three-dimensional density field from a set of one-dimensional spectra, interpolating the information between the lines of sight. We show that our algorithm provides unbiased mass profiles of clusters, becoming an alternative for estimating cluster masses complementary to weak lensing or X-ray observations. The algorithm employs a Hamiltonian Monte Carlo method to generate realizations of initial and evolved density fields and the three-dimensional large-scale flow, revealing the cosmic dynamics at high redshift. The method correctly handles multi-modal parameter distributions, which allow constraining the physics of the intergalactic medium (IGM) with high accuracy. We performed several tests using realistic simulated quasar spectra to test and validate our method. Our results show that detailed and physically plausible inference of three-dimensional large-scale structures at high redshift has become feasible.

astro-ph.CO↗

arXiv2019

Systematic-free inference of the cosmic matter density field from SDSS3-BOSS data

We perform an analysis of the three-dimensional cosmic matter density field traced by galaxies of the SDSS-III/BOSS galaxy sample. The systematic-free nature of this analysis is confirmed by two elements: the successful cross-correlation with the gravitational lensing observations derived from Planck 2018 data and the absence of bias at scales $k \simeq 10^{-3}-10^{-2}h$ Mpc$^{-1}$ in the a posteriori power spectrum of recovered initial conditions. Our analysis builds upon our algorithm for Bayesian Origin Reconstruction from Galaxies (BORG) and uses a physical model of cosmic structure formation to infer physically meaningful cosmic structures and their corresponding dynamics from deep galaxy observations. Our approach accounts for redshift-space distortions and light-cone effects inherent to deep observations. We also apply detailed corrections to account for known and unknown foreground contaminations, selection effects and galaxy biases. We obtain maps of residual, so far unexplained, systematic effects in the spectroscopic data of SDSS-III/BOSS. Our results show that unbiased and physically plausible models of the cosmic large scale structure can be obtained from present and next-generation galaxy surveys.

astro-ph.CO↗

arXiv2019

Neural physical engines for inferring the halo mass distribution function

An ambitious goal in cosmology is to forward-model the observed distribution of galaxies in the nearby Universe today from the initial conditions of large-scale structures. For practical reasons, the spatial resolution at which this can be done is necessarily limited. Consequently, one needs a mapping between the density of dark matter averaged over ~Mpc scales, and the distribution of dark matter halos (used as a proxy for galaxies) in the same region. Here we demonstrate a method for determining the halo mass distribution function by learning the tracer bias between density fields and halo catalogues using a neural bias model. The method is based on the Bayesian analysis of simple, physically motivated, neural network-like architectures, which we denote as neural physical engines, and neural density estimation. As a result, we are able to sample the initial phases of the dark matter density field whilst inferring the parameters describing the halo mass distribution function, providing a fully Bayesian interpretation of both the initial dark matter density distribution and the neural bias model. We successfully run an upgraded BORG inference using our new likelihood and neural bias model with halo catalogues derived from full N-body simulations. We notice orders of magnitude improvement in modelling compared to classical biasing techniques.

astro-ph.CO↗

arXiv2019

Painting halos from cosmic density fields of dark matter with physically motivated neural networks

We present a novel halo painting network that learns to map approximate 3D dark matter fields to realistic halo distributions. This map is provided via a physically motivated network with which we can learn the non-trivial local relation between dark matter density field and halo distributions without relying on a physical model. Unlike other generative or regressive models, a well motivated prior and simple physical principles allow us to train the mapping network quickly and with relatively little data. In learning to paint halo distributions from computationally cheap, analytical and non-linear density fields, we bypass the need for full particle mesh simulations and halo finding algorithms. Furthermore, by design, our halo painting network needs only local patches of dark matter density to predict the halos, and as such, it can predict the 3D halo distribution for any arbitrary simulation box size. Our neural network can be trained using small simulations and used to predict large halo distributions, as long as the resolutions are equivalent. We evaluate our model's ability to generate 3D halo count distributions which reproduce, to a high degree, summary statistics such as the power spectrum and bispectrum, of the input or reference realizations.

astro-ph.CO↗

arXiv2019

Explicit Bayesian treatment of unknown foreground contaminations in galaxy surveys

The treatment of unknown foreground contaminations will be one of the major challenges for galaxy clustering analyses of coming decadal surveys. These data contaminations introduce erroneous large-scale effects in recovered power spectra and inferred dark matter density fields. In this work, we present an effective solution to this problem in the form of a robust likelihood designed to account for effects due to unknown foreground and target contaminations. Conceptually, this robust likelihood marginalizes over the unknown large-scale contamination amplitudes. We showcase the effectiveness of this novel likelihood via an application to a mock SDSS-III data set subject to dust extinction contamination. In order to illustrate the performance of our proposed likelihood, we infer the underlying dark-matter density field and reconstruct the matter power spectrum, being maximally agnostic about the foregrounds. The results are compared to those of an analysis with a standard Poissonian likelihood, as typically used in modern large-scale structure analyses. While the standard Poissonian analysis yields excessive power for large-scale modes and introduces an overall bias in the power spectrum, our likelihood provides unbiased estimates of the matter power spectrum over the entire range of Fourier modes considered in this work. Further, we demonstrate that our approach accurately accounts for and corrects the effects of unknown foreground contaminations when inferring three-dimensional density fields. Robust likelihood approaches, as presented in this work, will be crucial to control unknown systematic error and maximize the outcome of the decadal surveys.

astro-ph.CO↗

arXiv2019

A rigorous EFT-based forward model for large-scale structure

Conventional approaches to cosmology inference from galaxy redshift surveys are based on n-point functions, which are under rigorous perturbative control on sufficiently large scales. Here, we present an alternative approach, which employs a likelihood at the level of the galaxy density field. By integrating out small-scale modes based on effective-field theory arguments, we prove that this likelihood is under perturbative control if certain specific conditions are met. We further show that the information captured by this likelihood is equivalent to the combination of the next-to-leading order galaxy power spectrum, leading-order bispectrum, and BAO reconstruction. Combined with MCMC sampling and MAP optimization techniques, our results allow for fully Bayesian cosmology inference from large-scale structure that is under perturbative control. We illustrate this via a first demonstration of unbiased cosmology inference from nonlinear large-scale structure using this likelihood. In particular, we show unbiased estimates of the power spectrum normalization $σ_{8}$ from a catalog of simulated dark matter halos, where nonlinear information is crucial in breaking the $b_{1} - σ_{8}$ degeneracy.

astro-ph.CO↗

arXiv2018

The Fifth Force in the Local Cosmic Web

Extensions of the standard models of particle physics and cosmology often lead to long-range fifth forces with properties dependent on gravitational environment. Fifth forces on astrophysical scales are best studied in the cosmic web where perturbation theory breaks down. We present constraints on chameleon- and symmetron-screened fifth forces with Yukawa coupling and megaparsec range -- as well as unscreened fifth forces with differential coupling to galactic mass components -- by searching for the displacements they predict between galaxies' stars and gas. Taking data from the Alfalfa HI survey, identifying galaxies' gravitational environments with the maps of Desmond et al. (2018a) and forward-modelling with a Bayesian likelihood framework, we set upper bounds on fifth-force strength relative to Newtonian gravity from $ΔG/G_N < \text{few} \: \times 10^{-4}$ for range $λ_C = 50$ Mpc, to $ΔG/G_N \lesssim 0.1$ for $λ_C = 500$ kpc. In $f(R)$ gravity this requires $f_{R0} < \text{few} \: \times \: 10^{-8}$. The analogous bounds without screening are $ΔG/G_N < \text{few} \: \times 10^{-4}$ and $ΔG/G_N < \text{few} \times 10^{-3}$. These are the tightest and among the only fifth-force constraints on galaxy scales. We show how our results may be strengthened with future survey data and identify the key features of an observational programme for furthering fifth-force tests beyond the Solar System.

astro-ph.CO↗

arXiv2018

Cosmological inference from Bayesian forward modelling of deep galaxy redshift surveys

We present a large-scale Bayesian inference framework to constrain cosmological parameters using galaxy redshift surveys, via an application of the Alcock-Paczyński (AP) test. Our physical model of the non-linearly evolved density field, as probed by galaxy surveys, employs Lagrangian perturbation theory (LPT) to connect Gaussian initial conditions to the final density field, followed by a coordinate transformation to obtain the redshift space representation for comparison with data. We generate realizations of primordial and present-day matter fluctuations given a set of observations. This hierarchical approach encodes a novel AP test, extracting several orders of magnitude more information from the cosmological expansion compared to classical approaches, to infer cosmological parameters and jointly reconstruct the underlying 3D dark matter density field. The novelty of this AP test lies in constraining the comoving-redshift transformation to infer the appropriate cosmology which yields isotropic correlations of the galaxy density field, with the underlying assumption relying purely on the cosmological principle. Such an AP test does not rely explicitly on modelling the full statistics of the field. We verify in depth via simulations that this renders our test robust to model misspecification. This leads to another crucial advantage, namely that the cosmological parameters exhibit extremely weak dependence on the currently unresolved phenomenon of galaxy bias, thereby circumventing a potentially key limitation. This is consequently among the first methods to extract a large fraction of information from statistics other than that of direct density contrast correlations, without being sensitive to the amplitude of density fluctuations. We perform several statistical efficiency and consistency tests on a mock galaxy catalogue, using the SDSS-III survey as template.

astro-ph.CO↗

arXiv2018

Fifth force constraints from galaxy warps

Intra-galaxy signals contain a wealth of information on fundamental physics, both the dark sector and the nature of gravity. While so far largely unexplored, such probes are set to rise dramatically in importance as upcoming surveys provide data of unprecedented quantity and quality on galaxy structure and dynamics. In this paper, we use warping of stellar disks to test the chameleon- or symmetron-screened fifth forces which generically arise when new fields couple to matter. We take r-band images of mostly late-type galaxies from the Nasa Sloan Atlas and develop an automated algorithm to quantify the degree of U-shaped warping they exhibit. We then forward-model the warp signal as a function of fifth-force strength $ΔG/G_N$ and range $λ_C$, and the gravitational environments and internal properties of the galaxies, including full propagation of the non-Gaussian uncertainties. Convolving this fifth-force likelihood function with a Gaussian describing astrophysical and observational noise and then constraining $ΔG/G_N$ and $λ_C$ by Markov Chain Monte Carlo, we find the overall likelihood to be significant increased ($Δ\log(\mathcal{L}) \simeq 20$) by adding a screened fifth force with $λ_C \simeq 2$ Mpc, $ΔG/G_N \simeq 0.01$. The variation of $Δ\log(\mathcal{L})$ with $λ_C$ is quantitatively as expected from the correlation of the magnitude of the fifth-force field with the force's range, and a similar model without screening achieves no increase in likelihood over the General Relativistic case $ΔG=0$. Although these results are in good agreement with a previous analysis of the same model using offsets between galaxies' stellar and gas mass centroids (Desmond et al. 2018), we caution that the effects of confounding baryonic and dark matter physics must be thoroughly investigated for the results of the inference to be unambiguous.

astro-ph.CO↗

arXiv2018

Fifth force constraints from the separation of galaxy mass components

One of the most common consequences of extensions to the standard models of particle physics or cosmology is the emergence of a fifth force. While generic fifth forces are tightly constrained at Solar System scales and below, they may escape detection by means of a screening mechanism which effectively removes them in dense environments. We constrain the strength $ΔG/G_N$ and range $λ_C$ of a chameleon- or symmetron-screened fifth force with Yukawa coupling -- as well as an unscreened fifth force with differential coupling to galactic mass components -- by searching for the displacement it predicts between galaxies' stellar and gas mass centroids. Taking data from the Alfalfa HI survey, identifying galaxies' gravitational environments with the maps of Desmond et al. (2018a) and forward-modelling with a Bayesian likelihood framework, we find $6.6σ$ evidence for $ΔG>0$ at $λ_C \simeq 2$ Mpc, with $ΔG/G_N = 0.025$ at maximum-likelihood. A similar fifth-force model without screening gives no increase in likelihood over the case $ΔG = 0$ for any $λ_C$. Although we validate these results by several methods, we do not claim screened modified gravity to provide the only possible explanation for the signal: this would require knowing that "galaxy formation" physics could not be responsible. We show also the results of a more conservative -- though less well motivated -- noise model which yields only upper limits on $ΔG/G_N$, ranging from $\sim10^{-1}$ for $λ_C \simeq 0.5$ Mpc to $\sim \: \text{few} \times 10^{-4}$ at $λ_C \simeq 50$ Mpc. We show how these constraints may be improved by future galaxy surveys and identify the key features of an observational programme for directly constraining fifth forces on galactic scales. This paper provides a complete description of the analysis summarised in Desmond et al. (2018b).

astro-ph.CO↗

arXiv2018

Automatic physical inference with information maximising neural networks

Compressing large data sets to a manageable number of summaries that are informative about the underlying parameters vastly simplifies both frequentist and Bayesian inference. When only simulations are available, these summaries are typically chosen heuristically, so they may inadvertently miss important information. We introduce a simulation-based machine learning technique that trains artificial neural networks to find non-linear functionals of data that maximise Fisher information: information maximising neural networks (IMNNs). In test cases where the posterior can be derived exactly, likelihood-free inference based on automatically derived IMNN summaries produces nearly exact posteriors, showing that these summaries are good approximations to sufficient statistics. In a series of numerical examples of increasing complexity and astrophysical relevance we show that IMNNs are robustly capable of automatically finding optimal, non-linear summaries of the data even in cases where linear compression fails: inferring the variance of Gaussian signal in the presence of noise; inferring cosmological parameters from mock simulations of the Lyman-α forest in quasar spectra; and inferring frequency-domain parameters from LISA-like detections of gravitational waveforms. In this final case, the IMNN summary outperforms linear data compression by avoiding the introduction of spurious likelihood maxima. We anticipate that the automatic physical inference method described in this paper will be essential to obtain both accurate and precise cosmological parameter estimates from complex and large astronomical data sets, including those from LSST and Euclid.

astro-ph.IM↗

arXiv2018

Physical Bayesian modelling of the non-linear matter distribution: new insights into the Nearby Universe

Accurate analyses of present and next-generation galaxy surveys require new ways to handle effects of non-linear gravitational structure formation in data. To address these needs we present an extension of our previously developed algorithm for Bayesian Origin Reconstruction from Galaxies to analyse matter clustering at non-linear scales in observations. This is achieved by incorporating a numerical particle mesh model of structure formation into our Bayesian inference framework. The algorithm simultaneously infers the 3D primordial matter fluctuations from which present non-linear observations formed and provides reconstructions of velocity fields and structure formation histories. The physical forward modelling approach automatically accounts for non-Gaussian features in evolved matter density fields and addresses the redshift space distortion problem associated with peculiar motions of galaxies. Our algorithm employs a hierarchical Bayes approach to jointly account for observational effects, such as galaxy biases, selection effects, and noise. Corresponding parameters are marginalized out via a sophisticated Markov Chain Monte Carlo approach relying on a combination of a multiple block sampling framework and a Hamiltonian Monte Carlo sampler. We demonstrate the performance of the method by applying it to the 2M++ galaxy compilation, tracing the matter distribution of the Nearby Universe. We show accurate and detailed inferences of the 3D non-linear dark matter distribution of the Nearby Universe. As exemplified in the case of the Coma cluster, we provide mass estimates that are compatible with those obtained from weak lensing and X-ray observations. For the first time, we reconstruct the vorticity of the non-linear velocity field from observations. In summary, our method provides plausible and detailed inferences of dark matter and velocity fields of our cosmic neighbourhood.

astro-ph.CO↗

arXiv2018

Optimal and fast E/B separation with a dual messenger field

We adapt our recently proposed dual messenger algorithm for spin field reconstruction and showcase its efficiency and effectiveness in Wiener filtering polarized cosmic microwave background (CMB) maps. Unlike conventional preconditioned conjugate gradient (PCG) solvers, our preconditioner-free technique can deal with high-resolution joint temperature and polarization maps with inhomogeneous noise distributions and arbitrary mask geometries with relative ease. Various convergence diagnostics illustrate the high quality of the dual messenger reconstruction. In contrast, the PCG implementation fails to converge to a reasonable solution for the specific problem considered. The implementation of the dual messenger method is straightforward and guarantees numerical stability and convergence. We show how the algorithm can be modified to generate fluctuation maps, which, combined with the Wiener filter solution, yield unbiased constrained signal realizations, consistent with observed data. This algorithm presents a pathway to exact global analyses of high-resolution and high-sensitivity CMB data for a statistically optimal separation of E and B modes. It is therefore relevant for current and next-generation CMB experiments, in the quest for the elusive primordial B-mode signal.

astro-ph.CO↗

arXiv2018

Imprints of the large-scale structure on AGN formation and evolution

Black hole masses are found to correlate with several global properties of their host galaxies, suggesting that black holes and galaxies have an intertwined evolution and that active galactic nuclei (AGN) have a significant impact on galaxy evolution. Since the large-scale environment can also affect AGN, this work studies how their formation and properties depend on the environment. We have used a reconstructed three-dimensional high-resolution density field obtained from a Bayesian large-scale structure reconstruction method applied to the 2M++ galaxy sample. A web-type classification relying on the shear tensor is used to identify different structures on the cosmic web, defining voids, sheets, filaments, and clusters. We confirm that the environmental density affects the AGN formation and their properties. We found that the AGN abundance is equivalent to the galaxy abundance, indicating that active and inactive galaxies reside in similar dark matter halos. However, occurrence rates are different for each spectral type and accretion rate. These differences are consistent with the AGN evolutionary sequence suggested by previous authors, Seyferts and Transition objects transforming into LINERs (Low-Ionization Nuclear Emission Line Regions), the weaker counterpart of Seyferts. We conclud that AGN properties depend on the environmental density more than on the web-type. More powerful starbursts and younger stellar populations are found in high densities, where interactions and mergers are more likely. AGN hosts show smaller masses in clusters for Seyferts and Transition objects, which might be due to gas stripping. In voids, the AGN population is dominated by the most massive galaxy hosts.

astro-ph.GA↗