arXiv subjects

Yashar Hezaveh

Publications and source records attributed to Yashar Hezaveh.

At least 55 records · Page 3Linked to original sources

arXiv2023

AstroPhot: Fitting Everything Everywhere All at Once in Astronomical Images

We present AstroPhot, a fast, powerful, and user-friendly Python based astronomical image photometry solver. AstroPhot incorporates automatic differentiation and GPU (or parallel CPU) acceleration, powered by the machine learning library PyTorch. Everything: AstroPhot can fit models for sky, stars, galaxies, PSFs, and more in a principled Chi^2 forward optimization, recovering Bayesian posterior information and covariance of all parameters. Everywhere: AstroPhot can optimize forward models on CPU or GPU; across images that are large, multi-band, multi-epoch, rotated, dithered, and more. All at once: The models are optimized together, thus handling overlapping objects and including the covariance between parameters (including PSF and galaxy parameters). A number of optimization algorithms are available including Levenberg-Marquardt, Gradient descent, and No-U-Turn MCMC sampling. With an object-oriented user interface, AstroPhot makes it easy to quickly extract detailed information from complex astronomical data for individual images or large survey programs. This paper outlines novel features of the AstroPhot code and compares it to other popular astronomical image modeling software. AstroPhot is open-source, fully Python based, and freely accessible here: https://github.com/Autostronomy/AstroPhot

astro-ph.IM↗

arXiv2023

TEMPLATES: Characterization of a Merger in the Dusty Lensing SPT0418-47 System

We present JWST and ALMA results for the lensing system SPT0418-47, which includes a strongly-lensed, dusty star-forming galaxy at redshift z=4.225 and an associated multiply-imaged companion. JWST NIRCam and MIRI imaging observations presented in this paper were acquired as part of the Early Release Science program Targeting Extremely Magnified Panchromatic Lensed Arcs and Their Extended Star Formation (TEMPLATES). This data set provides robust, mutiwavelength detection of stellar light in both the main (SPT0418A) and companion (SPT0418B) galaxies, while the ALMA detection of [C II] emission confirms that SPT0418B lies at the same redshift as SPT0418A. From a source plane reconstruction, we infer that the projected physical separation of the two galaxies is $4.42\pm 0.05$ kpc. We derive total magnifications of $μ=29.5\pm1.2$ and $μ=4.2\pm 0.9$ for SPT0418A and SPT0418B, respectively. We use both CIGALE and PROSPECTOR to derive stellar masses. The stellar mass ratio of SPT0418A and SPT0418B is approximately 4 to 1 ($4.5\pm 1.0$ for CIGALE and $4.2^{+1.9}_{-1.6}$ for PROSPECTOR). We also see evidence of extended structure associated with SPT0418A in the lensing reconstruction that is suggestive of a tidal feature. Interestingly, the star formation rates and stellar masses of both galaxies are consistent with the main sequence of star-forming galaxies at this epoch, indicating that this ongoing interaction has not noticeably elevated the star formation levels.

astro-ph.GA↗

arXiv2023

Sampling-Based Accuracy Testing of Posterior Estimators for General Inference

Parameter inference, i.e. inferring the posterior distribution of the parameters of a statistical model given some data, is a central problem to many scientific disciplines. Generative models can be used as an alternative to Markov Chain Monte Carlo methods for conducting posterior inference, both in likelihood-based and simulation-based problems. However, assessing the accuracy of posteriors encoded in generative models is not straightforward. In this paper, we introduce `Tests of Accuracy with Random Points' (TARP) coverage testing as a method to estimate coverage probabilities of generative posterior estimators. Our method differs from previously-existing coverage-based methods, which require posterior evaluations. We prove that our approach is necessary and sufficient to show that a posterior estimator is accurate. We demonstrate the method on a variety of synthetic examples, and show that TARP can be used to test the results of posterior inference analyses in high-dimensional spaces. We also show that our method can detect inaccurate inferences in cases where existing methods fail.

stat.ML↗

arXiv2023

Pixelated Reconstruction of Foreground Density and Background Surface Brightness in Gravitational Lensing Systems using Recurrent Inference Machines

Modeling strong gravitational lenses in order to quantify the distortions in the images of background sources and to reconstruct the mass density in the foreground lenses has been a difficult computational challenge. As the quality of gravitational lens images increases, the task of fully exploiting the information they contain becomes computationally and algorithmically more difficult. In this work, we use a neural network based on the Recurrent Inference Machine (RIM) to simultaneously reconstruct an undistorted image of the background source and the lens mass density distribution as pixelated maps. The method iteratively reconstructs the model parameters (the image of the source and a pixelated density map) by learning the process of optimizing the likelihood given the data using the physical model (a ray-tracing simulation), regularized by a prior implicitly learned by the neural network through its training data. When compared to more traditional parametric models, the proposed method is significantly more expressive and can reconstruct complex mass distributions, which we demonstrate by using realistic lensing galaxies taken from the IllustrisTNG cosmological hydrodynamic simulation.

astro-ph.IM↗

arXiv2023

Posterior Sampling of the Initial Conditions of the Universe from Non-linear Large Scale Structures using Score-Based Generative Models

Reconstructing the initial conditions of the universe is a key problem in cosmology. Methods based on simulating the forward evolution of the universe have provided a way to infer initial conditions consistent with present-day observations. However, due to the high complexity of the inference problem, these methods either fail to sample a distribution of possible initial density fields or require significant approximations in the simulation model to be tractable, potentially leading to biased results. In this work, we propose the use of score-based generative models to sample realizations of the early universe given present-day observations. We infer the initial density field of full high-resolution dark matter N-body simulations from the present-day density field and verify the quality of produced samples compared to the ground truth based on summary statistics. The proposed method is capable of providing plausible realizations of the early universe density field from the initial conditions posterior distribution marginalized over cosmological parameters and can sample orders of magnitude faster than current state-of-the-art methods.

astro-ph.CO↗

arXiv2023

Beyond Gaussian Noise: A Generalized Approach to Likelihood Analysis with non-Gaussian Noise

Likelihood analysis is typically limited to normally distributed noise due to the difficulty of determining the probability density function of complex, high-dimensional, non-Gaussian, and anisotropic noise. This is a major limitation for precision measurements in many domains of science, including astrophysics, for example, for the analysis of the Cosmic Microwave Background, gravitational waves, gravitational lensing, and exoplanets. This work presents Score-based LIkelihood Characterization (SLIC), a framework that resolves this issue by building a data-driven noise model using a set of noise realizations from observations. We show that the approach produces unbiased and precise likelihoods even in the presence of highly non-Gaussian correlated and spatially varying noise. We use diffusion generative models to estimate the gradient of the probability density of noise with respect to data elements. In combination with the Jacobian of the physical model of the signal, we use Langevin sampling to produce independent samples from the unbiased likelihood. We demonstrate the effectiveness of the method using real data from the Hubble Space Telescope and James Webb Space Telescope.

astro-ph.IM↗

arXiv2023

Extended Lyman-$α$ emission towards the SPT2349-56 protocluster at $z=4.3$

Context. Deep spectroscopic surveys with the Atacama Large Millimeter/submillimeter Array (ALMA) have revealed that some of the brightest infrared sources in the sky correspond to concentrations of dusty star-forming galaxies (DSFG) at high redshift. Among these, the SPT2349-56 protocluster system at z = 4.304 is amongst the most extreme examples due to its high source density and integrated star formation rate. Aims. We conducted a deep Lyman-$α$ line emission survey around SPT2349-56 using the Multi-Unit Spectroscopic Explorer (MUSE) at Very Large Telescope (VLT) in order to characterize this uniquely dense environment. Methods. Taking advantage of the deep three-dimensional nature of this survey, we performed a sensitive search for Lyman-$α$ emitters (LAEs) toward the core and northern extension of the protocluster, which correspond to the brightest infrared regions in this field. Using a smoothed narrowband image extracted from the MUSE datacube around the protocluster redshift, we searched for possible extended structures. Results. We identify only three LAEs at z = 4.3 in this field, in concordance with expectations for blank-fields, and an extended Lyman-$α$ structure spatially associated with core of the protocluster. All the previously-identified DSFGs in this field are undetected in Lyman-$α$ emission, consistent with the conspicuous dust obscuration in these systems. We find an extended Lyman-$α$ structure, about $60 \times 60$ kpc$^{2}$ in size, and located 56 kpc west of the protocluster core. Three DSFGs coincide spatially with the location of this structure. We conclude that either the three co-spatial DSFGs or the protocluster core itself are feeding ionizing photons to the Lyman-$α$ structure.

astro-ph.GA↗

arXiv2022

A Framework for Obtaining Accurate Posteriors of Strong Gravitational Lensing Parameters with Flexible Priors and Implicit Likelihoods using Density Estimation

We report the application of implicit likelihood inference to the prediction of the macro-parameters of strong lensing systems with neural networks. This allows us to perform deep learning analysis of lensing systems within a well-defined Bayesian statistical framework to explicitly impose desired priors on lensing variables, to obtain accurate posteriors, and to guarantee convergence to the optimal posterior in the limit of perfect performance. We train neural networks to perform a regression task to produce point estimates of lensing parameters. We then interpret these estimates as compressed statistics in our inference setup and model their likelihood function using mixture density networks. We compare our results with those of approximate Bayesian neural networks, discuss their significance, and point to future directions. Based on a test set of 100,000 strong lensing simulations, our amortized model produces accurate posteriors for any arbitrary confidence interval, with a maximum percentage deviation of $1.4\%$ at $21.8\%$ confidence level, without the need for any added calibration procedure. In total, inferring 100,000 different posteriors takes a day on a single GPU, showing that the method scales well to the thousands of lenses expected to be discovered by upcoming sky surveys.

astro-ph.IM↗

arXiv2022

Posterior samples of source galaxies in strong gravitational lenses with score-based priors

Inferring accurate posteriors for high-dimensional representations of the brightness of gravitationally-lensed sources is a major challenge, in part due to the difficulties of accurately quantifying the priors. Here, we report the use of a score-based model to encode the prior for the inference of undistorted images of background galaxies. This model is trained on a set of high-resolution images of undistorted galaxies. By adding the likelihood score to the prior score and using a reverse-time stochastic differential equation solver, we obtain samples from the posterior. Our method produces independent posterior samples and models the data almost down to the noise level. We show how the balance between the likelihood and the prior meet our expectations in an experiment with out-of-distribution data.

astro-ph.IM↗

arXiv2022

Population-Level Inference of Strong Gravitational Lenses with Neural Network-Based Selection Correction

A new generation of sky surveys is poised to provide unprecedented volumes of data containing hundreds of thousands of new strong lensing systems in the coming years. Convolutional neural networks are currently the only state-of-the-art method that can handle the onslaught of data to discover and infer the parameters of individual systems. However, many important measurements that involve strong lensing require population-level inference of these systems. In this work, we propose a hierarchical inference framework that uses the inference of individual lensing systems in combination with the selection function to estimate population-level parameters. In particular, we show that it is possible to model the selection function of a CNN-based lens finder with a neural network classifier, enabling fast inference of population-level parameters without the need for expensive Monte Carlo simulations.

astro-ph.IM↗

arXiv2022

Pixelated Reconstruction of Gravitational Lenses using Recurrent Inference Machines

Modeling strong gravitational lenses in order to quantify the distortions in the images of background sources and to reconstruct the mass density in the foreground lenses has traditionally been a difficult computational challenge. As the quality of gravitational lens images increases, the task of fully exploiting the information they contain becomes computationally and algorithmically more difficult. In this work, we use a neural network based on the Recurrent Inference Machine (RIM) to simultaneously reconstruct an undistorted image of the background source and the lens mass density distribution as pixelated maps. The method we present iteratively reconstructs the model parameters (the source and density map pixels) by learning the process of optimization of their likelihood given the data using the physical model (a ray-tracing simulation), regularized by a prior implicitly learned by the neural network through its training data. When compared to more traditional parametric models, the proposed method is significantly more expressive and can reconstruct complex mass distributions, which we demonstrate by using realistic lensing galaxies taken from the cosmological hydrodynamic simulation IllustrisTNG.

astro-ph.IM↗

arXiv2022

Simulation-Based Inference of Strong Gravitational Lensing Parameters

In the coming years, a new generation of sky surveys, in particular, Euclid Space Telescope (2022), and the Rubin Observatory's Legacy Survey of Space and Time (LSST, 2023) will discover more than 200,000 new strong gravitational lenses, which represents an increase of more than two orders of magnitude compared to currently known sample sizes. Accurate and fast analysis of such large volumes of data under a statistical framework is therefore crucial for all sciences enabled by strong lensing. Here, we report on the application of simulation-based inference methods, in particular, density estimation techniques, to the predictions of the set of parameters of strong lensing systems from neural networks. This allows us to explicitly impose desired priors on lensing parameters, while guaranteeing convergence to the optimal posterior in the limit of perfect performance.

astro-ph.CO↗

arXiv2022

Rapid build-up of the stellar content in the protocluster core SPT2349$-$56 at $z\,{=}\,4.3$

The protocluster SPT2349$-$56 at $z\,{=}\,4.3$ contains one of the most actively star-forming cores known, yet constraints on the total stellar mass of this system are highly uncertain. We have therefore carried out deep optical and infrared observations of this system, probing rest-frame ultraviolet to infrared wavelengths. Using the positions of the spectroscopically-confirmed protocluster members, we identify counterparts and perform detailed source deblending, allowing us to fit spectral energy distributions in order to estimate stellar masses. We show that the galaxies in SPT2349$-$56 have stellar masses proportional to their high star-formation rates, consistent with other protocluster galaxies and field submillimetre galaxies (SMGs) around redshift 4. The galaxies in SPT2349$-$56 have on average lower molecular gas-to-stellar mass fractions and depletion timescales than field SMGs, although with considerable scatter. We construct the stellar-mass function for SPT2349$-$56 and compare it to the stellar-mass function of $z\,{=}\,1$ galaxy clusters, finding consistent shapes between the two. We measure rest-frame galaxy ultraviolet half-light radii from our HST-F160W imaging, finding that on average the galaxies in our sample are similar in size to typical star-forming galaxies at these redshifts. However, the brightest HST-detected galaxy in our sample, found near the luminosity-weighted centre of the protocluster core, remains unresolved at this wavelength. Hydrodynamical simulations predict that the core galaxies will quickly merge into a brightest cluster galaxy, thus our observations provide a direct view of the early formation mechanisms of this class of object.

astro-ph.GA↗

arXiv2021

Molecular Line Observations in Two Dusty Star-Forming Galaxies at z = 6.9

SPT0311-58 is the most massive infrared luminous system discovered so far during the Epoch of Reionization (EoR). In this paper, we present a detailed analysis of the molecular interstellar medium at z = 6.9, through high-resolution observations of the CO(6-5), CO(7-6), CO(10-9), [CI](2-1), and p-H2O(211-202) lines and dust continuum emission with the Atacama Large Millimeter/submillimeter Array (ALMA). The system consists of a pair of intensely star-forming gravitationally lensed galaxies (labelled West and East). The intrinsic far-infrared luminosity is (16 $\pm$ 4)$\times\rm 10^{12} \ \rm L_{\odot}$ in West and (27 $\pm$ 4)$\times\rm 10^{11} \ \rm L_{\odot}$ in East. We model the dust, CO, and [CI] using non-local thermodynamic equilibrium radiative transfer models and estimate the intrinsic gas mass to be (5.4 $\pm$ 3.4)$\times\rm 10^{11} \ \rm M_{\odot}$ in West and (3.1 $\pm$ 2.7)$\times\rm 10^{10} \ \rm M_{\odot}$ in East. We find that the CO spectral line energy distribution in West and East are typical of high-redshift sub-millimeter galaxies (SMGs). The CO-to-H2 conversion factor ($α_{CO}$) and the gas depletion time scales estimated from the model are consistent with the high-redshift SMGs in the literature within the uncertainties. We find no evidence of evolution of depletion time with redshift in SMGs at z > 3. This is the most detailed study of molecular gas content of a galaxy in the EoR to-date, with the most distant detection of H2O in a galaxy without any evidence for active galactic nuclei in the literature.

astro-ph.GA↗

arXiv2020

Overdensities of Submillimetre-Bright Sources around Candidate Protocluster Cores Selected from the South Pole Telescope Survey

We present APEX-LABOCA 870 micron observations of the fields surrounding the nine brightest, high-redshift, unlensed objects discovered in the South Pole Telescope's (SPT) 2500 square degrees survey. Initially seen as point sources by SPT's 1-arcmin beam, the 19-arcsec resolution of our new data enables us to deblend these objects and search for submillimetre (submm) sources in the surrounding fields. We find a total of 98 sources above a threshold of 3.7 sigma in the observed area of 1300 square arcminutes, where the bright central cores resolve into multiple components. After applying a radial cut to our LABOCA sources to achieve uniform sensitivity and angular size across each of the nine fields, we compute the cumulative and differential number counts and compare them to estimates of the background, finding a significant overdensity of approximately 10 at 14 mJy. The large overdensities of bright submm sources surrounding these fields suggest that they could be candidate protoclusters undergoing massive star-formation events. Photometric and spectroscopic redshifts of the unlensed central objects range from 3 to 7, implying a volume density of star-forming protoclusters of approximately 0.1 per giga-parsec cube. If the surrounding submm sources in these fields are at the same redshifts as the central objects, then the total star-formation rates of these candidate protoclusters reach 10,000 solar masses per year, making them much more active at these redshifts than what has been seen so far in both simulations and observations.

astro-ph.CO↗

arXiv2020

Megaparsec-scale structure around the proto-cluster core SPT2349$-$56 at $z\,{=}\,4.3$

We present an extensive ALMA spectroscopic follow-up programme of the $z\,{=}\,4.3$ structure SPT2349$-$56, one of the most actively star-forming proto-cluster cores known, to identify additional members using their [C{\sc ii}] 158\,$μ$m and \mbox{CO(4--3)} lines. In addition to robustly detecting the 14 previously published galaxies in this structure, we identify a further 15 associated galaxies at $z\,{=}\,4.3$, resolving 55$\,{\pm}\,$5\,per cent of the 870-$μ$m flux density at 0.5\,arcsec resolution compared to 21\,arcsec single-dish data. These galaxies are distributed into a central core containing 23 galaxies extending out to 300\,kpc in diameter, and a northern extension, offset from the core by 400\,kpc, containing three galaxies. We discovered three additional galaxies in a red {\it Herschel\/}-SPIRE source 1.5\,Mpc from the main structure, suggesting the existence of many other sources at the same redshift as SPT2349$-$56 that are not yet detected in the limited coverage of our data. An analysis of the velocity distribution of the central galaxies indicates that this region may be virialized with a mass of (9$\pm$5)$\,{\times}\,$10$^{12}$\,M$_{\odot}$, while the two offset galaxy groups are about 30 and 60\,per cent less massive and show significant velocity offsets from the central group. We calculate the [C{\sc ii}] and far-infrared number counts, and find evidence for a break in the [C{\sc ii}] luminosity function. We estimate the average SFR density within the region of SPT2349$-$56 containing single-dish emission (a proper diametre of 720\,kpc), assuming spherical symmetry, to be roughly 4$\,{\times}\,10^4$\,M$_{\odot}$\,yr$^{-1}$\,Mpc$^{-3}$; this may be an order of magnitude greater than the most extreme examples seen in simulations.

astro-ph.GA↗

arXiv2019

Spatially Resolved Water Emission from Gravitationally Lensed Dusty Star Forming Galaxies at z $\sim$ 3

Water ($\rm H_{2}O$), one of the most ubiquitous molecules in the universe, has bright millimeter-wave emission lines easily observed at high-redshift with the current generation of instruments. The low excitation transition of $\rm H_{2}O$, p$-$$\rm H_{2}O$(202 $-$ 111) ($ν_{rest}$ = 987.927 GHz) is known to trace the far-infrared (FIR) radiation field independent of the presence of active galactic nuclei (AGN) over many orders-of-magnitude in FIR luminosity (L$_{\rm FIR}$). This indicates that this transition arises mainly due to star formation. In this paper, we present spatially ($\sim$0.5 arcsec corresponding to $\sim$1 kiloparsec) and spectrally resolved ($\sim$100 kms$^{-1}$) observations of p$-$$\rm H_{2}O$(202 $-$ 111) in a sample of four strong gravitationally lensed high-redshift galaxies with the Atacama Large Millimeter/submillimeter Array (ALMA). In addition to increasing the sample of luminous ($ > $ $10^{12}$L$_{\odot}$) galaxies observed with $\rm H_{2}O$, this paper examines the L$_{\rm H_{2}O}$/L$_{\rm FIR}$ relation on resolved scales for the first time at high-redshift. We find that L$_{\rm H_{2}O}$ is correlated with L$_{\rm FIR}$ on both global and resolved kiloparsec scales within the galaxy in starbursts and AGN with average L$_{\rm H_{2}O}$/L$_{\rm FIR}$ =$2.76^{+2.15}_{-1.21}\times10^{-5}$. We find that the scatter in the observed L$_{\rm H_{2}O}$/L$_{\rm FIR}$ relation does not obviously correlate with the effective temperature of the dust spectral energy distribution (SED) or the molecular gas surface density. This is a first step in developing p$-$$\rm H_{2}O$(202 $-$ 111) as a resolved star formation rate (SFR) calibrator.

astro-ph.GA↗

arXiv2019

Imaging the Molecular Interstellar Medium in a Gravitationally Lensed Star-forming Galaxy at z=5.7

Aims: We present and study spatially resolved imaging obtained with the Atacama Large Millimeter/submillimeter Array (ALMA) of multiple $^{12}$CO($J=$6$-$5, 8$-$7 and 9$-$8) and two H$_2$O(2$_{02}-$1$_{11}$ and 2$_{11}-$2$_{02}$) emission lines and cold dust continuum toward the gravitationally lensed dusty star forming galaxy SPT0346-52 at z=$5.656$. Methods: Using a visibility-domain source-plane reconstruction we probe the structure and dynamics of the different components of the interstellar medium (ISM) in this galaxy down to scales of 1 kpc in the source plane. Results: Measurements of the intrinsic sizes of the different CO emission lines indicate that the higher J transitions trace more compact regions in the galaxy. Similarly, we find smaller dust continuum intrinsic sizes with decreasing wavelength, based on observations at rest-frame 130, 300 and 450$μ$m. The source shows significant velocity structure, and clear asymmetry where an elongated structure is observed in the source plane with significant variations in their reconstructed sizes. This could be attributed to a compact merger or turbulent disk rotation. The differences in velocity structure through the different line tracers, however, hint at the former scenario in agreement with previous [CII] line imaging results. Measurements of the CO line ratios and magnifications yield significant variations as a function of velocity, suggesting that modeling of the ISM using integrated values could be misinterpreted. Modeling of the ISM in SPT0346-52 based on delensed fluxes indicate a highly dense and warm medium, qualitatively similar to that observed in high redshift quasar hosts.

astro-ph.GA↗