arXiv subjects

Stefano Cavuoti

Publications and source records attributed to Stefano Cavuoti.

At least 37 records · Page 2Linked to original sources

arXiv2021

The LSST-DESC 3x2pt Tomography Optimization Challenge

This paper presents the results of the Rubin Observatory Dark Energy Science Collaboration (DESC) 3x2pt tomography challenge, which served as a first step toward optimizing the tomographic binning strategy for the main DESC analysis. The task of choosing an optimal tomographic binning scheme for a photometric survey is made particularly delicate in the context of a metacalibrated lensing catalogue, as only the photometry from the bands included in the metacalibration process (usually riz and potentially g) can be used in sample definition. The goal of the challenge was to collect and compare bin assignment strategies under various metrics of a standard 3x2pt cosmology analysis in a highly idealized setting to establish a baseline for realistically complex follow-up studies; in this preliminary study, we used two sets of cosmological simulations of galaxy redshifts and photometry under a simple noise model neglecting photometric outliers and variation in observing conditions, and contributed algorithms were provided with a representative and complete training set. We review and evaluate the entries to the challenge, finding that even from this limited photometry information, multiple algorithms can separate tomographic bins reasonably well, reaching figures-of-merit scores close to the attainable maximum. We further find that adding the g band to riz photometry improves metric performance by ~15% and that the optimal bin assignment strategy depends strongly on the science case: which figure-of-merit is to be optimized, and which observables (clustering, lensing, or both) are included.

astro-ph.IM↗

arXiv2021

Improving the reliability of photometric redshift with machine learning

In order to answer the open questions of modern cosmology and galaxy evolution theory, robust algorithms for calculating photometric redshifts (photo-z) for very large samples of galaxies are needed. Correct estimation of the various photo-z algorithms' performance requires attention to both the performance metrics and the data used for the estimation. In this work, we use the supervised machine learning algorithm MLPQNA to calculate photometric redshifts for the galaxies in the COSMOS2015 catalogue and the unsupervised Self-Organizing Maps (SOM) to determine the reliability of the resulting estimates. We find that for spec-z<1.2, photo-z predictions are on the same level of quality as SED fitting photo-z. We show that the SOM successfully detects unreliable spec-z that cause biases in the estimation of the photo-z algorithms' performance. Additionally, we use SOM to select the objects with reliable photo-z predictions. Our cleaning procedures allow to extract the subset of objects for which the quality of the final photo-z catalogs is improved by a factor of two, compared to the overall statistics.

astro-ph.IM↗

arXiv2021

The connection between star formation and supermassive Black Hole activity in the local Universe

We present a study of the active galactic nucleus (AGN) activity in the local Universe (z < 0.33) and its correlation with the host galaxy properties, derived from a Sloan Digital Sky Survey (SDSS DR8) sample with spectroscopic star-formation rate (SFR) and stellar mass ($\mathcal{M}_{\ast}$) determination. To quantify the level of AGN activity we used X-ray information from the XMM-Newton Serendipitous Source Catalogue (3XMM DR8). Applying multiwavelength AGN selection criteria (optical BPT-diagrams, X-ray/optical ratio etc) we found that 24% of the detected sources are efficiently-accreting AGN with moderate-to-high X-ray luminosity, which are twice as likely to be hosted by star-forming galaxies than by quiescent ones. The distribution of the specific Black Hole accretion rate (sBHAR, $λ_{\mathrm{sBHAR}}$) shows that nuclear activity in local, non-AGN dominated galaxies peaks at very low accretion rates ($-4 \lesssim \logλ_{\mathrm{sBHAR}} \lesssim -3$) in all stellar mass ranges. However, we observe systematically larger values of sBHAR for galaxies with active star-formation than for quiescent ones, as well as an increase of the mean $λ_{\mathrm{sBHAR}}$ with SFR for both star-forming and quiescent galaxies. These findings confirm the decreased level of AGN activity with cosmic time and are consistent with a scenario where both star-formation and AGN activity are fuelled by a common gas reservoir.

astro-ph.GA↗

arXiv2021

Photometric redshifts with machine learning, lights and shadows on a complex data science use case

The current role of data-driven science is constantly increasing its importance within Astrophysics, due to the huge amount of multi-wavelength data collected every day, characterized by complex and high-volume information requiring efficient and as much as possible automated exploration tools. Furthermore, to accomplish main and legacy science objectives of future or incoming large and deep survey projects, such as JWST, LSST and Euclid, a crucial role is played by an accurate estimation of photometric redshifts, whose knowledge would permit the detection and analysis of extended and peculiar sources by disentangling low-z from high-z sources and would contribute to solve the modern cosmological discrepancies. The recent photometric redshift data challenges, organized within several survey projects, like LSST and Euclid, pushed the exploitation of multi-wavelength and multi-dimensional data observed or ad hoc simulated to improve and optimize the photometric redshifts prediction and statistical characterization based on both SED template fitting and machine learning methodologies. But they also provided a new impetus in the investigation on hybrid and deep learning techniques, aimed at conjugating the positive peculiarities of different methodologies, thus optimizing the estimation accuracy and maximizing the photometric range coverage, particularly important in the high-z regime, where the spectroscopic ground truth is poorly available. In such a context we summarize what learned and proposed in more than a decade of research.

astro-ph.IM↗

arXiv2020

Rejection criteria based on outliers in the KiDS photometric redshifts and PDF distributions derived by machine learning

The Probability Density Function (PDF) provides an estimate of the photometric redshift (zphot) prediction error. It is crucial for current and future sky surveys, characterized by strict requirements on the zphot precision, reliability and completeness. The present work stands on the assumption that properly defined rejection criteria, capable of identifying and rejecting potential outliers, can increase the precision of zphot estimates and of their cumulative PDF, without sacrificing much in terms of completeness of the sample. We provide a way to assess rejection through proper cuts on the shape descriptors of a PDF, such as the width and the height of the maximum PDF's peak. In this work we tested these rejection criteria to galaxies with photometry extracted from the Kilo Degree Survey (KiDS) ESO Data Release 4, proving that such approach could lead to significant improvements to the zphot quality: e.g., for the clipped sample showing the best trade-off between precision and completeness, we achieve a reduction in outliers fraction of $\simeq 75\%$ and an improvement of $\simeq 6\%$ for NMAD, with respect to the original data set, preserving the $\simeq 93\%$ of its content.

astro-ph.IM↗

arXiv2020

Anomaly detection in Astrophysics: a comparison between unsupervised Deep and Machine Learning on KiDS data

Every field of Science is undergoing unprecedented changes in the discovery process, and Astronomy has been a main player in this transition since the beginning. The ongoing and future large and complex multi-messenger sky surveys impose a wide exploiting of robust and efficient automated methods to classify the observed structures and to detect and characterize peculiar and unexpected sources. We performed a preliminary experiment on KiDS DR4 data, by applying to the problem of anomaly detection two different unsupervised machine learning algorithms, considered as potentially promising methods to detect peculiar sources, a Disentangled Convolutional Autoencoder and an Unsupervised Random Forest. The former method, working directly on images, is considered potentially able to identify peculiar objects like interacting galaxies and gravitational lenses. The latter instead, working on catalogue data, could identify objects with unusual values of magnitudes and colours, which in turn could indicate the presence of singularities.

astro-ph.IM↗

arXiv2020

Comparison of outlier detection methods on astronomical image data

Among the many challenges posed by the huge data volumes produced by the new generation of astronomical instruments there is also the search for rare and peculiar objects. Unsupervised outlier detection algorithms may provide a viable solution. In this work we compare the performances of six methods: the Local Outlier Factor, Isolation Forest, k-means clustering, a measure of novelty, and both a normal and a convolutional autoencoder. These methods were applied to data extracted from SDSS stripe 82. After discussing the sensitivity of each method to its own set of hyperparameters, we combine the results from each method to rank the objects and produce a final list of outliers.

astro-ph.IM↗

arXiv2020

Building the largest spectroscopic sample of ultra-compact massive galaxies with the Kilo Degree Survey

Ultra-compact massive galaxies UCMGs, i.e. galaxies with stellar masses $M_{*} > 8 \times 10^{10} M_{\odot}$ and effective radii $R_{e} < 1.5$ kpc, are very rare systems, in particular at low and intermediate redshifts. Their origin as well as their number density across cosmic time are still under scrutiny, especially because of the paucity of spectroscopically confirmed samples. We have started a systematic census of UCMG candidates within the ESO Kilo Degree Survey, together with a large spectroscopic follow-up campaign to build the largest possible sample of confirmed UCMGs. This is the third paper of the series and the second based on the spectroscopic follow-up program. Here, we present photometrical and structural parameters of 33 new candidates at redshifts $0.15 \lesssim z \lesssim 0.5$ and confirm 19 of them as UCMGs, based on their nominal spectroscopically inferred $M_{*}$ and $R_{e}$. This corresponds to a success rate of $\sim 58\%$, nicely consistent with our previous findings. The addition of these 19 newly confirmed objects, allows us to fully assess the systematics on the system selection, and finally reduce the number density uncertainties. Moreover, putting together the results from our current and past observational campaigns and some literature data, we build the largest sample of UCMGs ever collected, comprising 92 spectroscopically confirmed objects at $0.1 \lesssim z \lesssim 0.5$. This number raises to 116, allowing for a $3σ$ tolerance on the $M_{*}$ and $R_{e}$ thresholds for the UCMG definition. For all these galaxies we have estimated the velocity dispersion values at the effective radii which have been used to derive a preliminary mass-velocity dispersion correlation.

astro-ph.GA↗

arXiv2019

Astroinformatics based search for globular clusters in the Fornax Deep Survey

In the last years, Astroinformatics has become a well defined paradigm for many fields of Astronomy. In this work we demonstrate the potential of a multidisciplinary approach to identify globular clusters (GCs) in the Fornax cluster of galaxies taking advantage of multi-band photometry produced by the VLT Survey Telescope using automatic self-adaptive methodologies. The data analyzed in this work consist of deep, multi-band, partially overlapping images centered on the core of the Fornax cluster. In this work we use a Neural-Gas model, a pure clustering machine learning methodology, to approach the GC detection, while a novel feature selection method ($Φ$LAB) is exploited to perform the parameter space analysis and optimization. We demonstrate that the use of an Astroinformatics based methodology is able to provide GC samples that are comparable, in terms of purity and completeness with those obtained using single band HST data (Brescia et al. 2012) and two approaches based respectively on a morpho-photometric (Cantiello et al. 2018b) and a PCA analysis (D'Abrusco et al. 2015) using the same data discussed in this work.

astro-ph.IM↗

arXiv2019

Stellar formation rates in galaxies using Machine Learning models

Global Stellar Formation Rates or SFRs are crucial to constrain theories of galaxy formation and evolution. SFR's are usually estimated via spectroscopic observations which require too much previous telescope time and therefore cannot match the needs of modern precision cosmology. We therefore propose a novel method to estimate SFRs for large samples of galaxies using a variety of supervised ML models.

astro-ph.IM↗

arXiv2018

Statistical analysis of probability density functions for photometric redshifts through the KiDS-ESO-DR3 galaxies

Despite the high accuracy of photometric redshifts (zphot) derived using Machine Learning (ML) methods, the quantification of errors through reliable and accurate Probability Density Functions (PDFs) is still an open problem. First, because it is difficult to accurately assess the contribution from different sources of errors, namely internal to the method itself and from the photometric features defining the available parameter space. Second, because the problem of defining a robust statistical method, always able to quantify and qualify the PDF estimation validity, is still an open issue. We present a comparison among PDFs obtained using three different methods on the same data set: two ML techniques, METAPHOR (Machine-learning Estimation Tool for Accurate PHOtometric Redshifts) and ANNz2, plus the spectral energy distribution template fitting method, BPZ. The photometric data were extracted from the KiDS (Kilo Degree Survey) ESO Data Release 3, while the spectroscopy was obtained from the GAMA (Galaxy and Mass Assembly) Data Release 2. The statistical evaluation of both individual and stacked PDFs was done through quantitative and qualitative estimators, including a dummy PDF, useful to verify whether different statistical estimators can correctly assess PDF quality. We conclude that, in order to quantify the reliability and accuracy of any zphot PDF method, a combined set of statistical estimators is required.

astro-ph.IM↗

arXiv2018

Data Deluge in Astrophysics: Photometric Redshifts as a Template Use Case

Astronomy has entered the big data era and Machine Learning based methods have found widespread use in a large variety of astronomical applications. This is demonstrated by the recent huge increase in the number of publications making use of this new approach. The usage of machine learning methods, however is still far from trivial and many problems still need to be solved. Using the evaluation of photometric redshifts as a case study, we outline the main problems and some ongoing efforts to solve them.

astro-ph.IM↗

arXiv2018

Return of the features. Efficient feature selection and interpretation for photometric redshifts

The explosion of data in recent years has generated an increasing need for new analysis techniques in order to extract knowledge from massive datasets. Machine learning has proved particularly useful to perform this task. Fully automatized methods have recently gathered great popularity, even though those methods often lack physical interpretability. In contrast, feature based approaches can provide both well-performing models and understandable causalities with respect to the correlations found between features and physical processes. Efficient feature selection is an essential tool to boost the performance of machine learning models. In this work, we propose a forward selection method in order to compute, evaluate, and characterize better performing features for regression and classification problems. Given the importance of photometric redshift estimation, we adopt it as our case study. We synthetically created 4,520 features by combining magnitudes, errors, radii, and ellipticities of quasars, taken from the SDSS. We apply a forward selection process, a recursive method in which a huge number of feature sets is tested through a kNN algorithm, leading to a tree of feature sets. The branches of the tree are then used to perform experiments with the random forest, in order to validate the best set with an alternative model. We demonstrate that the sets of features determined with our approach improve the performances of the regression models significantly when compared to the performance of the classic features from the literature. The found features are unexpected and surprising, being very different from the classic features. Therefore, a method to interpret some of the found features in a physical context is presented. The methodology described here is very general and can be used to improve the performance of machine learning models for any regression or classification task.

astro-ph.IM↗

arXiv2018

Neural Gas based classification of Globular Clusters

Within scientific and real life problems, classification is a typical case of extremely complex tasks in data-driven scenarios, especially if approached with traditional techniques. Machine Learning supervised and unsupervised paradigms, providing self-adaptive and semi-automatic methods, are able to navigate into large volumes of data characterized by a multi-dimensional parameter space, thus representing an ideal method to disentangle classes of objects in a reliable and efficient way. In Astrophysics, the identification of candidate Globular Clusters through deep, wide-field, single band images, is one of such cases where self-adaptive methods demonstrated a high performance and reliability. Here we experimented some variants of the known Neural Gas model, exploring both supervised and unsupervised paradigms of Machine Learning for the classification of Globular Clusters. Main scope of this work was to verify the possibility to improve the computational efficiency of the methods to solve complex data-driven problems, by exploiting the parallel programming with GPU framework. By using the astrophysical playground, the goal was to scientifically validate such kind of models for further applications extended to other contexts.

astro-ph.IM↗

arXiv2017

Astrophysical Data Analytics based on Neural Gas Models, using the Classification of Globular Clusters as Playground

In Astrophysics, the identification of candidate Globular Clusters through deep, wide-field, single band HST images, is a typical data analytics problem, where methods based on Machine Learning have revealed a high efficiency and reliability, demonstrating the capability to improve the traditional approaches. Here we experimented some variants of the known Neural Gas model, exploring both supervised and unsupervised paradigms of Machine Learning, on the classification of Globular Clusters, extracted from the NGC1399 HST data. Main focus of this work was to use a well-tested playground to scientifically validate such kind of models for further extended experiments in astrophysics and using other standard Machine Learning methods (for instance Random Forest and Multi Layer Perceptron neural network) for a comparison of performances in terms of purity and completeness.

astro-ph.IM↗

arXiv2017

Probability density estimation of photometric redshifts based on machine learning

Photometric redshifts (photo-z's) provide an alternative way to estimate the distances of large samples of galaxies and are therefore crucial to a large variety of cosmological problems. Among the various methods proposed over the years, supervised machine learning (ML) methods capable to interpolate the knowledge gained by means of spectroscopical data have proven to be very effective. METAPHOR (Machine-learning Estimation Tool for Accurate PHOtometric Redshifts) is a novel method designed to provide a reliable PDF (Probability density Function) of the error distribution of photometric redshifts predicted by ML methods. The method is implemented as a modular workflow, whose internal engine for photo-z estimation makes use of the MLPQNA neural network (Multi Layer Perceptron with Quasi Newton learning rule), with the possibility to easily replace the specific machine learning model chosen to predict photo-z's. After a short description of the software, we present a summary of results on public galaxy data (Sloan Digital Sky Survey - Data Release 9) and a comparison with a completely different method based on Spectral Energy Distribution (SED) template fitting.

astro-ph.IM↗

arXiv2017

$C^{3}$ : A Command-line Catalogue Cross-matching tool for modern astrophysical survey data

In the current data-driven science era, it is needed that data analysis techniques has to quickly evolve to face with data whose dimensions has increased up to the Petabyte scale. In particular, being modern astrophysics based on multi-wavelength data organized into large catalogues, it is crucial that the astronomical catalog cross-matching methods, strongly dependant from the catalogues size, must ensure efficiency, reliability and scalability. Furthermore, multi-band data are archived and reduced in different ways, so that the resulting catalogues may differ each other in formats, resolution, data structure, etc, thus requiring the highest generality of cross-matching features. We present $C^{3}$ (Command-line Catalogue Cross-match), a multi-platform application designed to efficiently cross-match massive catalogues from modern surveys. Conceived as a stand-alone command-line process or a module within generic data reduction/analysis pipeline, it provides the maximum flexibility, in terms of portability, configuration, coordinates and cross-matching types, ensuring high performance capabilities by using a multi-core parallel processing paradigm and a sky partitioning algorithm.

astro-ph.IM↗

arXiv2017

METAPHOR: Probability density estimation for machine learning based photometric redshifts

We present METAPHOR (Machine-learning Estimation Tool for Accurate PHOtometric Redshifts), a method able to provide a reliable PDF for photometric galaxy redshifts estimated through empirical techniques. METAPHOR is a modular workflow, mainly based on the MLPQNA neural network as internal engine to derive photometric galaxy redshifts, but giving the possibility to easily replace MLPQNA with any other method to predict photo-z's and their PDF. We present here the results about a validation test of the workflow on the galaxies from SDSS-DR9, showing also the universality of the method by replacing MLPQNA with KNN and Random Forest models. The validation test include also a comparison with the PDF's derived from a traditional SED template fitting method (Le Phare).

astro-ph.IM↗