Millions Of Galaxies Research Articles

Euclid preparation

The Euclid mission is expected to image millions of galaxies at high resolution, providing an extensive dataset with which to study galaxy evolution. Because galaxy morphology is both a fundamental parameter and one that is hard to determine for large samples, we investigate the application of deep learning in predicting the detailed morphologies of galaxies in Euclid using Zoobot, a convolutional neural network pretrained with 450 000 galaxies from the Galaxy Zoo project. We adapted Zoobot for use with emulated Euclid images generated based on Hubble Space Telescope COSMOS images and with labels provided by volunteers in the Galaxy Zoo: Hubble project. We experimented with different numbers of galaxies and various magnitude cuts during the training process. We demonstrate that the trained Zoobot model successfully measures detailed galaxy morphology in emulated Euclid images. It effectively predicts whether a galaxy has features and identifies and characterises various features, such as spiral arms, clumps, bars, discs, and central bulges. When compared to volunteer classifications, Zoobot achieves mean vote fraction deviations of less than 12% and an accuracy of above 91% for the confident volunteer classifications across most morphology types. However, the performance varies depending on the specific morphological class. For the global classes, such as disc or smooth galaxies, the mean deviations are less than 10%, with only 1000 training galaxies necessary to reach this performance. On the other hand, for more detailed structures and complex tasks, such as detecting and counting spiral arms or clumps, the deviations are slightly higher, of namely around 12% with 60 000 galaxies used for training. In order to enhance the performance on complex morphologies, we anticipate that a larger pool of labelled galaxies is needed, which could be obtained using crowd sourcing. We estimate that, with our model, the detailed morphology of approximately 800 million galaxies of the Euclid Wide Survey could be reliably measured and that approximately 230 million of these galaxies would display features. Finally, our findings imply that the model can be effectively adapted to new morphological labels. We demonstrate this adaptability by applying Zoobot to peculiar galaxies. In summary, our trained Zoobot CNN can readily predict morphological catalogues for Euclid images.

Systematic Effects in Galaxy-Galaxy Lensing with DESI

The Dark Energy Spectroscopic Instrument (DESI) survey will measure spectroscopic redshifts for millions of galaxies across roughly 14,000deg2 of the sky. Cross-correlating targets in the DESI survey with complementary imaging surveys allows us to measure and analyze shear distortions caused by gravitational lensing in unprecedented detail. In this work, we analyze a series of mock catalogs with ray-traced gravitational lensing and increasing sophistication to estimate systematic effects on galaxy-galaxy lensing estimators such as the tangential shear γt and the excess surface density ΔΣ. We employ mock catalogs tailored to the specific imaging surveys overlapping with the DESI survey: the Dark Energy Survey (DES), the Hyper Suprime-Cam (HSC) survey, and the Kilo-Degree Survey (KiDS). Among others, we find that fiber incompleteness can have significant effects on galaxy-galaxy lensing estimators but can be corrected effectively by up-weighting DESI targets with fibers by the inverse of the fiber assignment probability. Similarly, we show that intrinsic alignment and lens magnification are expected to be statistically significant given the precision forecasted for the DESI year-1 data set. Our study informs several analysis choices for upcoming cross-correlation studies of DESI with DES, HSC, and KiDS.

Characterizing tidal features around galaxies in cosmological simulations

ABSTRACT Tidal features provide signatures of recent mergers and offer a unique insight into the assembly history of galaxies. The Vera C. Rubin Observatory’s Legacy Survey of Space and Time (LSST) will enable an unprecedentedly large survey of tidal features around millions of galaxies. To decipher the contributions of mergers to galaxy evolution it will be necessary to compare the observed tidal features with theoretical predictions. Therefore, we use cosmological hydrodynamical simulations NewHorizon, eagle, IllustrisTNG, and Magneticum to produce LSST-like mock images of z ∼ 0 galaxies (z ∼ 0.2 for NewHorizon) with $M_{\scriptstyle \star ,\text{ 30 pkpc}}\ge 10^{9.5}$ M$_{\scriptstyle \odot }$. We perform a visual classification to identify tidal features and classify their morphology. We find broadly good agreement between the simulations regarding their overall tidal feature fractions: $f_{{\small NewHorizon}}=0.40\pm 0.06$, $f_{{\small EAGLE}}=0.37\pm 0.01$, $f_{{\small TNG}}=0.32\pm 0.01$, and $f_{{\small Magneticum}}=0.32\pm 0.01$, and their specific tidal feature fractions. Furthermore, we find excellent agreement regarding the trends of tidal feature fraction with stellar and halo mass. All simulations agree in predicting that the majority of central galaxies of groups and clusters exhibit at least one tidal feature, while the satellite members rarely show such features. This agreement suggests that gravity is the primary driver of the occurrence of visually identifiable tidal features in cosmological simulations, rather than subgrid physics or hydrodynamics. All predictions can be verified directly with LSST observations.

Survey Operations for the Dark Energy Spectroscopic Instrument

The Dark Energy Spectroscopic Instrument (DESI) survey is a spectroscopic survey of tens of millions of galaxies at 0 < z < 3.5 covering 14,000 sq. deg. of the sky. In its first 1.1 yr of survey operations, it has observed more than 14 million galaxies and 4 million stars. We describe the processes that govern DESI’s observations of the 15,000 fields composing the survey. This includes the planning of each night’s observations in the afternoon; automatic selection of fields to observe during the night; real-time assessment of field completeness on the basis of observing conditions during each exposure; reduction, redshifting, and quality assurance of each field of targets in the morning following observation; and updates to the list of future targets to observe on the basis of these results. We also compare the performance of the survey with historical expectations and find good agreement. Simulations of the weather and of DESI observations using the real field-selection algorithm show good agreement with the actual observations. After accounting for major unplanned shutdowns, the dark time survey is progressing about 7% faster than forecast, which is good agreement given approximations made in the simulations.

Euclid: Cosmology forecasts from the void-galaxy cross-correlation function with reconstruction

We have investigated the cosmological constraints that can be expected from measurement of the cross-correlation of galaxies with cosmic voids identified in the Euclid spectroscopic survey, which will include spectroscopic information for tens of millions of galaxies over 15 000 deg2 of the sky in the redshift range 0.9 ≤ z &lt; 1.8. We have done this using simulated measurements obtained from the Flagship mock catalogue, the official Euclid mock that closely matches the expected properties of the spectroscopic dataset. To mitigate anisotropic selection-bias effects, we have used a velocity field reconstruction method to remove large-scale redshift-space distortions from the galaxy field before void-finding. This allowed us to accurately model contributions to the observed anisotropy of the cross-correlation function arising from galaxy velocities around voids as well as from the Alcock–Paczynski effect, and we studied the dependence of constraints on the efficiency of reconstruction. We find that Euclid voids will be able to constrain the ratio of the transverse comoving distance DM and Hubble distance DH to a relative precision of about 0.3%, and the growth rate fσ8 to a precision of between 5% and 8% in each of the four redshift bins covering the full redshift range. In the standard cosmological model, this translates to a statistical uncertainty ΔΩm = ±0.0028 on the matter density parameter from voids, which is better than what can be achieved from either Euclid galaxy clustering and weak lensing individually. We also find that voids alone can measure the dark energy equation of state to a 6% precision.

Systems Engineering Approach for the SPHEREx Telescope Mission

AbstractThe Spectro‐Photometer for the History of the Universe, Epoch of Reionization and Ices Explorer (SPHEREx) was selected in 2019 as a NASA Astrophysics Medium Class Explorer mission to perform the first infrared all‐sky spectral survey. The mission consists of a single wide field of view telescope mounted on top of a Ball Configurable Platform (BCP) spacecraft bus. Over the course of its two‐year mission, the SPHEREx observatory will perform four all‐sky surveys and provide spectroscopic measurements of hundreds of millions of galaxies and a diversity of astronomical phenomena. SPHEREx is currently in phase C, post‐Critical Design Review (CDR) and is in the process of building and testing hardware in preparation for observatory Assembly, Test, and Integration (AI&amp;T) in early 2024. The observatory is expected to launch on a Falcon 9 rocket no earlier than February 2025.Project level systems engineering support is provided by the Jet Propulsion Laboratory (JPL) and oversees the technical development of the observatory and its ground operations. Systems engineers are staffed throughout all the systems and act as a key interface between the mission's many partners. This paper focuses on the systems engineering approach for the SPHEREx mission including its architecture, verification and validation methodology, and its risk and resource management approach.

Neural Stellar Population Synthesis Emulator for the DESI PROVABGS

The Probabilistic Value-Added Bright Galaxy Survey (PROVABGS) catalog will provide the posterior distributions of physical properties of >10 million DESI Bright Galaxy Survey galaxies. Each posterior distribution will be inferred from joint Bayesian modeling of observed photometry and spectroscopy using Markov Chain Monte Carlo sampling and the Hahn et al. stellar population synthesis (SPS) model. To make this computationally feasible, PROVABGS will use a neural emulator for the SPS model to accelerate the posterior inference. In this work, we present how we construct the emulator using the Alsing et al. approach and verify that it can be used to accurately infer galaxy properties. We confirm that the emulator is in excellent agreement with the original SPS model with ≪1% error and is 100× faster. In addition, we demonstrate that the posteriors of galaxy properties derived using the emulator are also in excellent agreement with those inferred using the original model. The neural emulator presented in this work is essential in bypassing the computational challenge posed in constructing the PROVABGS catalog. Furthermore, it demonstrates the advantages of emulation for scaling sophisticated analyses to millions of galaxies.

The PAU Survey and Euclid: Improving broadband photometric redshifts with multi-task learning

Current and future imaging surveys require photometric redshifts (photo-zs) to be estimated for millions of galaxies. Improving the photo-z quality is a major challenge but is needed to advance our understanding of cosmology. In this paper we explore how the synergies between narrow-band photometric data and large imaging surveys can be exploited to improve broadband photometric redshifts. We used a multi-task learning (MTL) network to improve broadband photo-z estimates by simultaneously predicting the broadband photo-z and the narrow-band photometry from the broadband photometry. The narrow-band photometry is only required in the training field, which also enables better photo-z predictions for the galaxies without narrow-band photometry in the wide field. This technique was tested with data from the Physics of the Accelerating Universe Survey (PAUS) in the COSMOS field. We find that the method predicts photo-zs that are 13% more precise down to magnitude iAB &lt; 23; the outlier rate is also 40% lower when compared to the baseline network. Furthermore, MTL reduces the photo-z bias for high-redshift galaxies, improving the redshift distributions for tomographic bins with z &gt; 1. Applying this technique to deeper samples is crucial for future surveys such as Euclid or LSST. For simulated data, training on a sample with iAB &lt; 23, the method reduces the photo-z scatter by 16% for all galaxies with iAB &lt; 25. We also studied the effects of extending the training sample with photometric galaxies using PAUS high-precision photo-zs, which reduces the photo-z scatter by 20% in the COSMOS field.

Commentary: Astronomy from space after the JWST

Commentary: Astronomy from space after the <i>JWST</i>

Target Selection and Validation of DESI Luminous Red Galaxies

The Dark Energy Spectroscopic Instrument (DESI) is carrying out a five-year survey that aims to measure the redshifts of tens of millions of galaxies and quasars, including 8 million luminous red galaxies (LRGs) in the redshift range 0.4 < z ≲ 1.0. Here we present the selection of the DESI LRG sample and assess its spectroscopic performance using data from Survey Validation (SV) and the first two months of the Main Survey. The DESI LRG sample, selected using g, r, z, and W1 photometry from the DESI Legacy Imaging Surveys, is highly robust against imaging systematics. The sample has a target density of 605 deg−2 and a comoving number density of 5 × 10−4 h 3 Mpc−3 in 0.4 < z < 0.8; this is a significantly higher density than previous LRG surveys (such as SDSS, BOSS, and eBOSS) while also extending to z ∼ 1. After applying a bright star veto mask developed for the sample, 98.9% of the observed LRG targets yield confident redshifts (with a catastrophic failure rate of 0.2% in the confident redshifts), and only 0.5% of the LRG targets are stellar contamination. The LRG redshift efficiency varies with source brightness and effective exposure time, and we present a simple model that accurately characterizes this dependence. In the appendices, we describe the extended LRG samples observed during SV.

The Hα and [O iii] λ5007 Luminosity Functions of 1.2 < z < 1.9 Emission-line Galaxies from Hubble Space Telescope (HST) Grism Spectroscopy

Euclid and the Roman Space Telescope (Roman) will soon use grism spectroscopy to detect millions of galaxies via their Hα and [O iii] λ5007 emission. To better constrain the expected galaxy counts from these instruments, we use a vetted sample of 4239 emission-line galaxies from the 3D Hubble Space Telescope survey to measure the Hα and [O iii] λ5007 luminosity functions between 1.16 < z < 1.90; this sample is ∼4 times larger than previous studies at this redshift. We find very good agreement with previous measurements for Hα, but for [O iii], we predict a higher number of intermediate-luminosity galaxies than from previous works. We find that, for both lines, the characteristic luminosity, , increases monotonically with redshift, and use the Hα luminosity function to calculate the epoch’s cosmic star formation rate density. We find that Hα-visible galaxies account for ∼81% of the epoch’s total star formation rate, and this value changes very little over the 1.16 < z < 1.56 redshift range. Finally, we derive the surface density of galaxies as a function of limiting flux and find that previous predictions for galaxy counts for the Euclid Wide Survey are unchanged, but there may be more [O iii] galaxies in the Roman High Latitude Survey than previously estimated.

The Dawes Review 10: The impact of deep learning for the analysis of galaxy surveys

AbstractThe amount and complexity of data delivered by modern galaxy surveys has been steadily increasing over the past years. New facilities will soon provide imaging and spectra of hundreds of millions of galaxies. Extracting coherent scientific information from these large and multi-modal data sets remains an open issue for the community and data-driven approaches such as deep learning have rapidly emerged as a potentially powerful solution to some long lasting challenges. This enthusiasm is reflected in an unprecedented exponential growth of publications using neural networks, which have gone from a handful of works in 2015 to an average of one paper per week in 2021 in the area of galaxy surveys. Half a decade after the first published work in astronomy mentioning deep learning, and shortly before new big data sets such as Euclid and LSST start becoming available, we believe it is timely to review what has been the real impact of this new technology in the field and its potential to solve key challenges raised by the size and complexity of the new datasets. The purpose of this review is thus two-fold. We first aim at summarising, in a common document, the main applications of deep learning for galaxy surveys that have emerged so far. We then extract the major achievements and lessons learned and highlight key open questions and limitations, which in our opinion, will require particular attention in the coming years. Overall, state-of-the-art deep learning methods are rapidly adopted by the astronomical community, reflecting a democratisation of these methods. This review shows that the majority of works using deep learning up to date are oriented to computer vision tasks (e.g. classification, segmentation). This is also the domain of application where deep learning has brought the most important breakthroughs so far. However, we also report that the applications are becoming more diverse and deep learning is used for estimating galaxy properties, identifying outliers or constraining the cosmological model. Most of these works remain at the exploratory level though which could partially explain the limited impact in terms of citations. Some common challenges will most likely need to be addressed before moving to the next phase of massive deployment of deep learning in the processing of future surveys; for example, uncertainty quantification, interpretability, data labelling and domain shift issues from training with simulations, which constitutes a common practice in astronomy.

Galaxy Spectra Neural Networks (GaSNets). I. Searching for Strong Lens Candidates in eBOSS Spectra Using Deep Learning

With the advent of new spectroscopic surveys from ground and space, observing up to hundreds of millions of galaxies, spectra classification will become overwhelming for standard analysis techniques. To prepare for this challenge, we introduce a family of deep learning tools to classify features in one-dimensional spectra. As the first application of these Galaxy Spectra neural Networks (GaSNets), we focus on tools specialized in identifying emission lines from strongly lensed star-forming galaxies in the eBOSS spectra. We first discuss the training and testing of these networks and define a threshold probability, P L , of 95% for the high-quality event detection. Then, using a previous set of spectroscopically selected strong lenses from eBOSS, confirmed with the Hubble Space Telescope (HST), we estimate a completeness of ∼80% as the fraction of lenses recovered above the adopted P L . We finally apply the GaSNets to ∼1.3M eBOSS spectra to collect the first list of ∼430 new high-quality candidates identified with deep learning from spectroscopy and visually graded as highly probable real events. A preliminary check against ground-based observations tentatively shows that this sample has a confirmation rate of 38%, in line with previous samples selected with standard (no deep learning) classification tools and confirmed by the HST. This first test shows that machine learning can be efficiently extended to feature recognition in the wavelength space, which will be crucial for future surveys like 4MOST, DESI, Euclid, and the China Space Station Telescope.

Impact of image persistence in theRoman Space TelescopeHigh-Latitude Survey

ABSTRACTThe High-Latitude Survey of the Nancy Grace Roman Space Telescope is expected to measure the positions and shapes of hundreds of millions of galaxies in an area of 2220 deg2. This survey will provide high-quality weak lensing data with unprecedented systematic control. The Roman Space Telescope will survey the sky in near-infrared (NIR) bands using Teledyne H4RG HgCdTe photodiode arrays. These NIR arrays exhibit an effect called persistence: charges that are trapped in the photodiodes during earlier exposures are gradually released into later exposures, leading to contamination of the images and potentially to errors in measured galaxy properties such as fluxes and shapes. In this work, we use image simulations that incorporate the persistence effect to study its impact on galaxy shape measurements and weak lensing signals. No significant spatial correlations are found between the galaxy shape changes induced by persistence. On the scales of interest for weak lensing cosmology, the effect of persistence on the weak lensing correlation function is about two orders of magnitude lower than the Roman Space Telescope additive shear error budget, indicating that the persistence effect is expected to be a subdominant contributor to the systematic error budget for weak lensing with the Roman Space Telescope given its current design.

UNITSIM-Galaxies: data release and clustering of emission-line galaxies

ABSTRACT New surveys such as European Space Agencys (ESA’s) Euclid mission are planned to map with unprecedented precision the large-scale structure of the Universe by measuring the 3D positions of tens of millions of galaxies. It is necessary to develop theoretically modelled galaxy catalogues to estimate the expected performance and to optimize the analysis strategy of these surveys. We populate two pairs of (1 h−1 Gpc)3 volume dark matter-only simulations from the UNIT project with galaxies using the Semi-Analytic Galaxy Evolution semi-analytic model of galaxy formation, coupled to the photoionization model get_emlines to estimate their H α emission. These catalogues represent a unique suite that includes galaxy formation physics and – thanks to the fixed-pair technique used – an effective volume of $\sim\!(5\,h^{-1}\,\rm {Gpc})^3$, which is several times larger than the Euclid survey. We present the performance of these data and create five additional emission-line galaxy (ELG) catalogues by applying a dust-attenuation model as well as adjusting the flux threshold as a function of redshift in order to reproduce Euclid-forecast dN/dz values. As a first application, we study the abundance and clustering of those model H α ELGs: for scales greater than ∼5 h−1 Mpc, we find a scale-independent bias with a value of b ∼ 1 at redshift z ∼ 0.5, that can increase nearly linearly to b ∼ 4 at z ∼ 2, depending on the ELG catalogue. Model galaxy properties, including their emission-line fluxes (with and without dust extinction) are publicly available.

Euclidpreparation – XXIII. Derivation of galaxy physical properties with deep machine learning using mock fluxes andH-band images

ABSTRACTNext-generation telescopes, like Euclid, Rubin/LSST, and Roman, will open new windows on the Universe, allowing us to infer physical properties for tens of millions of galaxies. Machine-learning methods are increasingly becoming the most efficient tools to handle this enormous amount of data, because they are often faster and more accurate than traditional methods. We investigate how well redshifts, stellar masses, and star-formation rates (SFRs) can be measured with deep-learning algorithms for observed galaxies within data mimicking the Euclid and Rubin/LSST surveys. We find that deep-learning neural networks and convolutional neural networks (CNNs), which are dependent on the parameter space of the training sample, perform well in measuring the properties of these galaxies and have a better accuracy than methods based on spectral energy distribution fitting. CNNs allow the processing of multiband magnitudes together with $H_{\scriptscriptstyle \rm E}$-band images. We find that the estimates of stellar masses improve with the use of an image, but those of redshift and SFR do not. Our best results are deriving (i) the redshift within a normalized error of &amp;lt;0.15 for 99.9 ${{\ \rm per\ cent}}$ of the galaxies with signal-to-noise ratio &amp;gt;3 in the $H_{\scriptscriptstyle \rm E}$ band; (ii) the stellar mass within a factor of two ($\sim\!0.3 \rm \ dex$) for 99.5 ${{\ \rm per\ cent}}$ of the considered galaxies; and (iii) the SFR within a factor of two ($\sim\!0.3 \rm \ dex$) for $\sim\!70{{\ \rm per\ cent}}$ of the sample. We discuss the implications of our work for application to surveys as well as how measurements of these galaxy parameters can be improved with deep learning.

Euclidpreparation

Photometric redshifts (photo-zs) are one of the main ingredients in the analysis of cosmological probes. Their accuracy particularly affects the results of the analyses of galaxy clustering with photometrically selected galaxies (GCph) and weak lensing. In the next decade, space missions such asEuclidwill collect precise and accurate photometric measurements for millions of galaxies. These data should be complemented with upcoming ground-based observations to derive precise and accurate photo-zs. In this article we explore how the tomographic redshift binning and depth of ground-based observations will affect the cosmological constraints expected from theEuclidmission. We focus on GCphand extend the study to include galaxy-galaxy lensing (GGL). We add a layer of complexity to the analysis by simulating several realistic photo-zdistributions based on theEuclidConsortium Flagship simulation and using a machine learning photo-zalgorithm. We then use the Fisher matrix formalism together with these galaxy samples to study the cosmological constraining power as a function of redshift binning, survey depth, and photo-zaccuracy. We find that bins with an equal width in redshift provide a higher figure of merit (FoM) than equipopulated bins and that increasing the number of redshift bins from ten to 13 improves the FoM by 35% and 15% for GCphand its combination with GGL, respectively. For GCph, an increase in the survey depth provides a higher FoM. However, when we include faint galaxies beyond the limit of the spectroscopic training data, the resulting FoM decreases because of the spurious photo-zs. When combining GCphand GGL, the number density of the sample, which is set by the survey depth, is the main factor driving the variations in the FoM. Adding galaxies at faint magnitudes and high redshift increases the FoM, even when they are beyond the spectroscopic limit, since the number density increase compensates for the photo-zdegradation in this case. We conclude that there is more information that can be extracted beyond the nominal ten tomographic redshift bins ofEuclidand that we should be cautious when adding faint galaxies into our sample since they can degrade the cosmological constraints.

Photometric redshift estimation with a convolutional neural network: NetZ

Galaxy redshifts are a key characteristic for nearly all extragalactic studies. Since spectroscopic redshifts require additional telescope and human resources, millions of galaxies are known without spectroscopic redshifts. Therefore, it is crucial to have methods for estimating the redshift of a galaxy based on its photometric properties, the so-called photo-z. We have developed NetZ, a new method using a convolutional neural network (CNN) to predict the photo-z based on galaxy images, in contrast to previous methods that often used only the integrated photometry of galaxies without their images. We use data from the Hyper Suprime-Cam Subaru Strategic Program (HSC SSP) in five different filters as the training data. The network over the whole redshift range between 0 and 4 performs well overall and especially in the high-z range, where it fares better than other methods on the same data. We obtained a precision |zpred − zref| of σ = 0.12 (68% confidence interval) with a CNN working for all galaxy types averaged over all galaxies in the redshift range of 0 to ∼4. We carried out a comparison with a network trained on point-like sources, highlighting the importance of morphological information for our redshift estimation. By limiting the scope to smaller redshift ranges or to luminous red galaxies, we find a further notable improvement. We have published more than 34 million new photo-z values predicted with NetZ. This shows that the new method is very simple and swift in application, and, importantly, it covers a wide redshift range that is limited only by the available training data. It is broadly applicable, particularly with regard to upcoming surveys such as the Rubin Observatory Legacy Survey of Space and Time, which will provide images of billions of galaxies with similar image quality as HSC. Our HSC photo-z estimates are also beneficial to the Euclid survey, given the overlap in the footprints of the HSC and Euclid.

Measuring Stellar Masses of Emission-line Galaxies at 1.2 &lt; z &lt; 1.9

Abstract The accurate measurement of stellar masses over a wide range of galaxy properties is essential for better constraining models of galaxy evolution. Emission-line galaxies (ELGs) tend to have better redshift estimates than continuum-selected objects, and have been shown to span a large range of physical properties, including stellar mass. Using data from the 3D-HST Treasury program, in this work, we construct a carefully vetted sample of 4350 ELGs at redshifts of 1.16 &lt; z &lt; 1.90. We combine the 3D-HST emission-line fluxes with far-UV through near-IR photometry, and use the MCSED spectral energy distribution fitting code to constrain the galaxies’ physical parameters, such as their star-formation rate and stellar masses. Our sample is consistent with the z ∼ 2 mass–metallicity relation. More importantly, we show that there is a simple, but tight correlation between stellar mass and absolute magnitude in a near-IR filter, which should prove particularly useful in terms of the rapid calculation of accurate stellar masses for millions of galaxies in upcoming missions such as Euclid, and the Nancy Grace Roman Space Telescope.

A self-supervised, physics-aware, Bayesian neural network architecture for modelling galaxy emission-line kinematics

ABSTRACT In the upcoming decades, large facilities, such as the SKA, will provide resolved observations of the kinematics of millions of galaxies. In order to assist in the timely exploitation of these vast data sets, we explore the use of a self-supervised, physics-aware neural network capable of Bayesian kinematic modelling of galaxies. We demonstrate the network’s ability to model the kinematics of cold gas in galaxies with an emphasis on recovering physical parameters and accompanying modelling errors. The model is able to recover rotation curves, inclinations and disc scale lengths for both CO and H i data which match well with those found in the literature. The model is also able to provide modelling errors over learned parameters, thanks to the application of quasi-Bayesian Monte Carlo dropout. This work shows the promising use of machine learning, and in particular, self-supervised neural networks, in the context of kinematically modelling galaxies. This work represents the first steps in applying such models for kinematic fitting and we propose that variants of our model would seem especially suitable for enabling emission-line science from upcoming surveys with e.g. the SKA, allowing fast exploitation of these large data sets.