YOLO-CL: Galaxy cluster detection in the SDSS with deep machine learning
Kirill Grishin, Simona Mei, Stéphane Ilic
备注：
14 pages, 12 fig Ilic co-conceived and developed YOLO-CL, applied it to SDSS images, and produced the plots for the completeness and purity analysis (2019-2021). Grishin reproduced Ilic's work and compared the final YOLO-CL cluster catalog to the other cluster finders (2021-2022). Mei co-conceived YOLO-CL, supervised the work and was the main writer of the paper text
(Abridged) Galaxy clusters are a powerful probe of cosmological models. Next generation large-scale optical and infrared surveys will reach unprecedented depths over large areas and require highly complete and pure cluster catalogs, with a well defined selection function. We have developed a new cluster detection algorithm YOLO-CL, which is a modified version of the state-of-the-art object detection deep convolutional network YOLO, optimized for the detection of galaxy clusters. We trained YOLO-CL on color images of the redMaPPer cluster detections in the SDSS. We find that YOLO-CL detects $95-98\%$ of the redMaPPer clusters, with a purity of $95-98\%$ calculated by applying the network to SDSS blank fields. When compared to the MCXC2021 X-ray catalog in the SDSS footprint,YOLO-CL is more complete then redMaPPer, which means that the neural network improved the cluster detection efficiency of its training sample. The YOLO-CL selection function is approximately constant with redshift, with respect to the MCXC2021 cluster mean X-ray surface brightness. YOLO-CL shows high performance when compared to traditional detection algorithms applied to SDSS. Deep learning networks benefit from a strong advantage over traditional galaxy cluster detection techniques because they do not need galaxy photometric and photometric redshift catalogs. This eliminates systematic uncertainties that can be introduced during source detection, and photometry and photometric redshift measurements. Our results show that YOLO-CL is an efficient alternative to traditional cluster detection methods. In general, this work shows that it is worth exploring the performance of deep convolution networks for future cosmological cluster surveys, such as the Rubin/LSST, Euclid or the Roman Space Telescope surveys.
We present the data release of the Uchuu-SDSS galaxies: a set of 32 high-fidelity galaxy lightcones constructed from the large Uchuu 2.1 trillion particle $N$-body simulation using Planck cosmology. We adopt subhalo abundance matching to populate the Uchuu-box halo catalogues with SDSS galaxy luminosities. These cubic box galaxy catalogues generated at several redshifts are combined to create the set of lightcones with redshift-evolving galaxy properties. The Uchuu-SDSS galaxy lightcones are built to reproduce the footprint and statistical properties of the SDSS main galaxy survey, along with …
We test the Emergent Gravity(EG) theory using the galaxy-galaxy lensing technique based on SDSS DR7 data. In the EG scenario, we do not expect color dependence of the galaxy sample in the 'apparent dark matter' predicted by EG, which is exerted only by the baryonic mass. If the baryonic mass is similar, then the predicted lensing profiles from the baryonic mass should be similar according to the EG, regardless of the color of the galaxy sample. We use the stellar mass of the galaxy as a proxy of its baryonic mass. We divide our galaxy sample into 5 stellar mass bins, and further classify them …
Using the largest single-survey sample of Type Ia supernovae (SNe Ia) to date, we study the relationship between properties of SNe Ia and those of their host galaxies, focusing primarily on correlations with Hubble residuals (HR). Our sample consists of 345 photometrically-classified or spectroscopically-confirmed SNeIa discovered as part of the SDSS-II Supernova Survey (SDSS-SNS). This analysis utilizes host-galaxy spectroscopy obtained during the SDSS-I/II spectroscopic survey and from an ancillary program on the SDSS-III Baryon Oscillation Spectroscopic Survey (BOSS) that obtained spectra f…
This paper presents the identification of galaxy clusters from the photometric redshift catalog based on three imaging surveys of SCUSS, SDSS, and unWISE. By applying a fast clustering algorithm, we obtain a total of 19,610 clusters in the redshift range of $0.05 < z < 0.65$ over a sky area of about 3,700 deg$^2$ in the south Galactic gap. Monte Carlo simulations show that the false detection rate is about 8.9\%. The redshift uncertainty is estimated to be about 0.013. The mass and richness of detected clusters are derived through the calibration based on the measurements of X-ray emission and…
The counts-in-cells (CIC) galaxy probability distribution depends on both the dark matter clustering amplitude $σ_8$ and the galaxy bias $b$. We present a theory for the CIC distribution based on a previous prescription of the underlying dark matter distribution and a linear volume transformation to redshift space. We show that, unlike the power spectrum, the CIC distribution breaks the degeneracy between $σ_8$ and $b$ on scales large enough that both bias and redshift distortions are still linear; thus we obtain a simultaneous fit for both parameters. We first validate the technique on the Mi…
We present new results on the relationship between central galaxies and dark matter haloes inferred from observations of satellite kinematics in the Sloan Digital Sky Survey (SDSS) DR7. We employ an updated analysis framework that includes detailed mock catalogues to model observational effects in SDSS. Our results constrain the colour-dependent conditional luminosity function (CLF) of dark matter haloes, as well as the radial profile of satellite galaxies. Confirming previous results, we find that red central galaxies live in more massive haloes than blue galaxies at fixed luminosity. Additio…
Context. Friends-of-friends algorithms are a common tool to detect galaxy groups and clusters in large survey data. In order to be as precise as possible, they have to be carefully calibrated using mock catalogues. Aims. We create an accurate and robust description of the matter distribution in the local Universe using the most up-to-date available data. This will provide the input for a specific cosmological test planned as follow-up to this work, and will be useful for general extragalactic and cosmological research. Methods. We created a set of galaxy group catalogues based on the 2MRS and …
Based on galaxies from the Sloan Digital Sky Survey (SDSS) and subhalos in the corresponding reconstructed region from the constrained simulation of ELUCID, we study the alignment of central galaxies relative to their host groups in the group catalog, as well as the alignment relative to the corresponding subhalos in the ELUCID simulation. Galaxies in observation are matched to dark matter subhalos in the ELUCID simulation using a novel neighborhood abundance matching method. In observation, the major axes of galaxies are found to be preferentially aligned to the major axes of their host group…
Groups in the Millennium Simulation and in SDSS DR7
比较了Millennium N-body模拟和Sloan数字巡天第七次数据发布（SDSS DR 7）星系和星系群星表，研究了星系群的性质和星系群中星系的分布。我们使用Tago等人分析SDSS数据时使用的相同的朋友的朋友方法，为千年半分析星系表构建模拟星系群星表。我们详细分析了集团光度，集团丰富度，维里半径，集团的大小和它们的均方根速度为四个体积有限的样品从观测和模拟。我们的结果表明，在所有的样本中，群的空间密度在一个数量级内一致，模拟星表与观测值之间具有相当好的一致性。对于更富有的群体，所有群体属性分布都具有相似的形状和幅度。对于星系对和小星系群，观测和模拟的星系群性质明显不同。此外，星系在小星系群中的空间分布是不同的：在星系群的外围，星系数分布并不一致，尽管在内部区域的一致性相对较好。分布的差异主要是由于SDSS样本的观测限制和半分析方法中产生过于紧凑和明亮的群体的问题。
We spectroscopically detected 838 likely, 448 probable, and 265 possible strong lens candidates within $\approx2$ million galaxy spectra contained within the extended Baryon Oscillation Spectroscopic Survey (eBOSS) from the sixteenth data release (DR16) of the Sloan Digital Sky Survey (SDSS). We apply the spectroscopic detection method of the Baryon Oscillation Spectroscopic Survey (BOSS) Emission-Line Lens Survey (BELLS) and add Gaussian fit information, grading, additional inspection observables, and additional inspection methods to improve our selection method. We observed 477 candidates wi…
We analyze a set of volume limited samples from the SDSS to study the dependence of galaxy colour on different environments of the cosmic web. We measure the local dimension of galaxies to determine the geometry of their embedding environments and find that filaments host a higher fraction of red galaxies than sheets at each luminosity. We repeat the analysis at a fixed density and recover the same trend which shows that galaxy colours depend on geometry of environments besides local density. At a fixed luminosity, the fraction of red galaxies in filaments and sheets increases with the extent …
Several recent observations using large datasets of galaxies showed non-random distribution of the spin directions of spiral galaxies, even when the galaxies are too far from each other to have gravitational interaction. Here, a dataset of $\sim8.7\cdot10^3$ spiral galaxies imaged by Hubble Space Telescope is used to test and profile a possible asymmetry between galaxy spin directions. The asymmetry between galaxies with opposite spin directions is compared to the asymmetry of galaxies from the Sloan Digital Sky Survey. The two datasets contain different galaxies at different redshift ranges, …
As the first paper in a series on the study of the galaxy-galaxy lensing from Sloan Digital Sky Survey Data Release 7 (SDSS DR7), we present our image processing pipeline that corrects the systematics primarily introduced by the Point Spread Function (PSF). Using this pipeline, we processed SDSS DR7 imaging data in $r$ band and generated a background galaxy catalog containing the shape information of each galaxy. Based on our own shape measurements of the galaxy images from SDSS DR7, we extract the galaxy-galaxy (GG) lensing signals around foreground spectroscopic galaxies binned in different …
The distribution of spin directions of $\sim6.4\cdot10^4$ SDSS spiral galaxies with spectra was examined, and compared to the distribution of $\sim3.3\cdot10^4$ Pan-STARRS galaxies. The analysis shows a statistically significant asymmetry between the number of SDSS galaxies with opposite spin directions, and the magnitude and direction of the asymmetry changes with the direction of observation and with the redshift. The redshift dependence shows that the distribution of the spin direction of SDSS galaxies becomes more asymmetric as the redshift gets higher. Fitting the distribution of the gala…
We use machine learning techniques to classify galaxy merger stages, which can unveil physical processes that drive the star formation and active galactic nucleus (AGN) activities during galaxy interaction. The sample contains 4,690 galaxies from the integral field spectroscopy survey SDSS-IV MaNGA, and can be separated to 1,060 merging galaxies and 3630 non-merging or unclassified galaxies. For the merger sample, there are 468, 125, 293, and 174 galaxies in (1) incoming pair phase, (2) first pericentric passage phase, (3) aproaching or just passing the apocenter, and (4) final coalescence pha…
We analyze the galaxy pairs in a volume limited sample ($M_r \leq -21$) from the SDSS to study the effects of galaxy interactions on the star formation rate and colour of galaxies in different environments. We study the star formation rate and colour of the paired galaxies as a function of projected separation and compare the results with their control samples matched in stellar mass, redshift and local density. We find that the major interactions significantly enhance the star formation rate in paired galaxies and turn them bluer with decreasing pair separation within $30$ kpc. The impact of …
Galaxy spectra are essential to probe the spatial distribution of galaxies in our Universe. To better interpret current and future spectroscopic galaxy redshift surveys, it is important to be able to simulate these data sets. We describe Uspec, a forward modeling tool to generate galaxy spectra taking into account some intrinsic galaxy properties as well as instrumental responses of a given telescope. The model for the intrinsic properties of the galaxy population, i.e., the luminosity functions, and size and spectral coefficients distribu- tions, was developed in an earlier work for broad-ban…
We present observational constraints on the galaxy-halo connection, focusing particularly on galaxy assembly bias, from a novel combination of counts-in-cylinders statistics, $P(N_{\rm{CIC}})$, with the standard measurements of the projected two-point correlation function, $w_{\rm{p}}(r_{\rm{p}})$, and number density, $n_{\rm{gal}}$, of galaxies. We measure $n_{\rm{gal}}$, $w_{\rm{p}}(r_{\rm{p}})$ and $P(N_{\rm{CIC}})$ for volume-limited, luminosity-threshold samples of galaxies selected from SDSS DR7, and use them to constrain halo occupation distribution (HOD) models, including a model in wh…
Using the combined resolving power of the Hubble Space Telescope and gravitational lensing, we resolve star-forming structures in a z~2.5 galaxy on scales much smaller than the usual kiloparsec diffraction limit of HST. SGAS J111020.0+645950.8 is a clumpy, star forming galaxy lensed by the galaxy cluster SDSS J1110+6459 at z = 0.659, with a total magnification ~30x across the entire arc. We use a hybrid parametric/non-parametric strong lensing mass model to compute the deflection and magnification of this giant arc, reconstruct the light distribution of the lensed galaxy in the source plane, a…
We present a machine learning (ML) approach for the prediction of galaxies' dark matter halo masses that achieves an improved performance over conventional methods. We train three ML algorithms (\texttt{XGBoost}, Random Forests, and neural network) to predict halo masses using a set of synthetic galaxy catalogues that are built by populating dark matter haloes in N-body simulations with galaxies, and that match both the clustering and the joint-distributions of properties of galaxies in the Sloan Digital Sky Survey (SDSS). We explore the correlation of different galaxy- and group-related prope…
Galaxy groups and clusters are the main tools used to test cosmological models and to study the environmental effect of galaxy formation. This work provides a catalogue of galaxy groups and clusters, as well as potentially merging systems based on the SDSS main galaxy survey. We identified galaxy groups and clusters using the modified friends-of-friends (FoF) group finder designed specifically for flux-limited galaxy surveys. The FoF group membership is refined by multimodality analysis to find subgroups and by using the group virial radius and escape velocity to expose unbound galaxies. We lo…
It has been shown, both in simulations and observationally, that the tidal field of a large galaxy can torque its satellites such that the major axis of satellite galaxies points towards their hosts. This so-called `shape alignment' has been observed in isolated Milky Way-like galaxies but not in `Local Group'-like pairs. In this study, we investigate the shape alignment of satellite galaxies in galaxy pairs similar to the Local Group identified in the Sloan Digital Sky Survey Data Release 13 (SDSS DR13). By stacking tens of thousands of satellite galaxies around primary galaxy pairs, we find …
We present cosmological parameter constraints based on a joint modeling of galaxy-lensing cross correlations and galaxy clustering measurements in the SDSS, marginalizing over small-scale modeling uncertainties using mock galaxy catalogs, without explicit modeling of galaxy bias. We show that our modeling method is robust to the impact of different choices for how galaxies occupy dark matter halos and to the impact of baryonic physics (at the $\sim2\%$ level in cosmological parameters) and test for the impact of covariance on the likelihood analysis and of the survey window function on the the…
We analyze a set of volume limited sample of galaxies from the SDSS to study the issue of cosmic homogeneity. We use the Renyi entropy of different order to probe the inhomogeneties present in the galaxy distributions. We also calculate the Renyi diveregence to quantify the deviations of the galaxy distribution from a homogeneous Poisson distribution on different length scales. We separately carry out the analysis using the overlapping spheres and the independent voxels. Our analysis suggests that the scale of homogeneity is underestimated in the smaller galaxy samples due to the suppression o…
We have carried a detailed analysis on the impact of cosmological redshift in the non-parametric approach to automated galaxy morphology classification. We artificially redshifted each galaxy from the EFIGI 4458 sample (re-centred at $z \sim 0$) simulating SDSS, DES, LSST and HST instruments set-ups over the range $0 < z < 1.5$. We then traced how the morphometry is degraded in each $z$ using MORFOMETRYKA. In the process, we re-sampled all catalogues to several resolutions and to a diverse signal-to-noise range, allowing us to understand the impact of image sampling and noise on our measuremen…
As a large-scale overdensity collapses, it affects the orientation and shape of galaxies that form, by exerting tidal shear along their axes. Therefore, the shapes of elliptical galaxies align with the tidal field of cosmic structures. This intrinsic alignment provides insights into galaxy formation and the primordial universe, complements late-time cosmological probes and constitutes a significant systematic effect for weak gravitational lensing observations. In the present study, we provide constraints on the linear alignment model using a fully Bayesian field-level approach, using galaxy sh…
As the second paper of a series on studying galaxy-galaxy lensing signals using the Sloan Digital Sky Survey Data Release 7 (SDSS DR7), we present our measurement and modelling of the lensing signals around groups of galaxies. We divide the groups into four halo mass bins, and measure the signals around four different halo-center tracers: brightest central galaxy (BCG), luminosity-weighted center, number-weighted center and X-ray peak position. For X-ray and SDSS DR7 cross identified groups, we further split the groups into low and high X-ray emission subsamples, both of which are assigned wit…
Comments on arXiv:1310.5067: "A response to `A self-consistent public catalogue of voids and superclusters in the SDSS Data Release 7 galaxy surveys' "
We present measurements of $E_G$, a probe of gravity from large-scale structure, using BOSS LOWZ and CMASS spectroscopic samples, with lensing measurements from SDSS (galaxy lensing) and Planck (CMB lensing). Using SDSS lensing and the BOSS LOWZ sample, we measure $\langle{E_G}\rangle=0.40^{+0.05}_{-0.04}$ (stat), $\pm 0.026$ (systematic), consistent with the predicted value from the Planck $Λ$CDM model, $E_G=0.46$. Using CMB lensing, we measure $\langle{E_G}\rangle=0.46^{+0.08}_{-0.09}$ (stat) for LOWZ (statistically consistent with galaxy lensing and Planck predictions) and $\langle{E_G}\ran…
We investigate the level of galaxy assembly bias in the Sloan Digital Sky Survey (SDSS) main galaxy sample using ELUCID, a state-of-the-art constrained simulation that accurately reconstructed the initial density perturbations within the SDSS volume. On top of the ELUCID haloes, we develop an extended HOD model that includes the assembly bias of central and satellite galaxies, parameterized as $\mathcal{Q}_\mathrm{cen}$ and $\mathcal{Q}_\mathrm{sat}$, respectively, to predict a suite of one- and two-point observables. In particular, our fiducial constraint employs the probability distribution …
We study galaxy distributions with Sloan Digital Sky Survey SDSS DR14 data and with simulations searching for variables that can constrain neutrino masses. To be specific, we consider the scenario of three active neutrino eigenstates with approximately the same mass, so $\sum m_ν= 3 m_ν$. Fitting the predictions of the $Λ$CDM model to the Sachs-Wolfe effect, $σ_8$, the galaxy power spectrum $P_\textrm{gal}(k)$, fluctuations of galaxy counts in spheres of radii ranging from $16/h$ to $128/h$ Mpc, BAO measurements, and $h = 0.678 \pm 0.009$, in various combinations, with free spectral index $n$,…
Constraint on the cosmological f(R) model from the multipole power spectrum of the SDSS luminous red galaxy sample and prospects for a future redshift survey
We present results from cross-correlating Planck CMB lensing maps with the Sloan Digital Sky Survey (SDSS) galaxy lensing shape catalog and BOSS galaxy catalogs. For galaxy position vs. CMB lensing cross-correlations, we measure the convergence signal around the galaxies in configuration space, using the BOSS LOWZ ($z\sim0.30$) and CMASS ($z\sim0.57$) samples. With fixed Planck 2015 cosmology, doing a joint fit with the galaxy clustering measurement, for the LOWZ (CMASS) sample we find a galaxy bias $b_g=1.75\pm0.04$ ($1.95\pm 0.02$) and galaxy-matter cross-correlation coefficient $r_{cc}=1.0\…
