Giuseppe Longo

Astroinformatics of galaxies and quasars: a new general method for photometric redshifts estimation

With the availability of the huge amounts of data produced by current and future large multiband photometric surveys, photometric redshifts have become a crucial tool for extragalactic astronomy and cosmology. In this paper we present a novel method, called Weak Gated Experts (WGE), which allows us to derive photometric redshifts through a combination of data mining techniques. The WGE, like many other machine learning techniques, is based on the exploitation of a spectroscopic knowledge base composed by sources for which a spectroscopic value of the redshift is available. This method achieves a variance σ^2(Δz) = 2.3 × 10^(−4) [σ^2(Δz) = 0.08, where Δz=z_(phot)−z_(spec)] for the reconstruction of the photometric redshifts for the optical galaxies from the Sloan Digital Sky Survey (SDSS) and for the optical quasars, respectively, while the root mean square (rms) of the Δz variable distributions for the two experiments is, respectively, equal to 0.021 and 0.35. The WGE provides also a mechanism for the estimation of the accuracy of each photometric redshift. We also present and discuss the catalogues obtained for the optical SDSS galaxies, for the optical candidate quasars extracted from the Data Release 7 of SDSS photometric data set (the sample of SDSS sources on which the accuracy of the reconstruction has been assessed is composed of bright sources, for a subset of which spectroscopic redshifts have been measured) and for optical SDSS candidate quasars observed by GALEX in the ultraviolet range. The WGE method exploits the new technological paradigm provided by the virtual observatory and the emerging field of astroinformatics.

Photometric classification of emission line galaxies with machine-learning methods

In this paper, we discuss an application of machine-learning-based methods to the identification of candidate active galactic nucleus (AGN) from optical survey data and to the automatic classification of AGNs in broad classes. We applied four different machine-learning algorithms, namely the Multi Layer Perceptron, trained, respectively, with the Conjugate Gradient, the Scaled Conjugate Gradient, the Quasi Newton learning rules and the Support Vector Machines, to tackle the problem of the classification of emission line galaxies in different classes, mainly AGNs versus non-AGNs, obtained using optical photometry in place of the diagnostics based on line intensity ratios which are classically used in the literature. Using the same photometric features, we discuss also the behaviour of the classifiers on finer AGN classification tasks, namely Seyfert I versus Seyfert II, and Seyfert versus LINER. Furthermore, we describe the algorithms employed, the samples of spectroscopically classified galaxies used to train the algorithms, the procedure followed to select the photometric parameters and the performances of our methods in terms of multiple statistical indicators. The results of the experiments show that the application of self-adaptive data mining algorithms trained on spectroscopic data sets and applied to carefully chosen photometric parameters represents a viable alternative to the classical methods that employ time-consuming spectroscopic observations.

The luminosity function of the NoSOCS galaxy cluster sample

We present the analysis of the luminosity function of a large sample of galaxy clusters from the Northern Sky Optical Cluster Survey, using latest data from the Sloan Digital Sky Survey. Our global luminosity function (down to M_r ≾ -16) does not show the presence of an upturn at faint magnitudes, while we do observe a strong dependence of its shape on both the richness and the clustercentric radius, with a brightening of M^* and an increase in the dwarf-to-giant galaxy ratio with the richness, indicating that more massive systems are more efficient in creating/retaining a population of dwarf satellites. This is observed within both physical (0.5R_(200)) and fixed (0.5 Mpc) apertures, suggesting that the trend is either due to a global effect, operating at all scales, or due to a local one but operating on even smaller scales. We further observe a decrease in the relative number of dwarf galaxies towards the cluster centre; this is most probably due to tidal collisions or the collisional disruption of the dwarfs since merging processes are inhibited by the high velocity dispersions in cluster cores and, furthermore, we do not observe a strong dependence of the bright end on the environment. nWe find an indication that the dwarf-to-giant ratio decreases with increasing redshift, within 0.07 ≤z < 0.2. We also measure a trend for the stronger suppression of faint galaxies (below M^*+ 2) with increasing redshift in poor systems, with respect to more massive ones, indicating that the evolutionary stage of less-massive galaxies depends more critically on the environment. nFinally, we point out that the luminosity function is far from universal; hence, the uncertainties introduced by the different methods used to build a composite function may partially explain the variety of faint-end slopes reported in the literature, as well as, in some cases, the presence of a faint-end upturn.

Variability-selected active galactic nuclei in the VST-SUDARE/VOICE survey of the COSMOS field

Optical variability has proven to be an effective way of detecting AGNs in imaging surveys, lasting from weeks to years. In the present work we test its use as a tool to identify AGNs in the VST multi-epoch survey of the COSMOS field, originally tailored to detect supernova events. We make use of the multi-wavelength data provided by other COSMOS surveys to discuss the reliability of the method and the nature of our AGN candidates. Our selection returns a sample of 83 AGN candidates; based on a number of diagnostics, we conclude that 67 of them are confirmed AGNs (81% purity), 12 are classified as supernovae, while the nature of the remaining 4 is unknown. For the subsample of AGNs with some spectroscopic classification, we find that Type 1 are prevalent (89%) compared to Type 2 AGNs (11%). Overall, our approach is able to retrieve on average 15% of all AGNs in the field identified by means of spectroscopic or X-ray classification, with a strong dependence on the source apparent magnitude. In particular, the completeness for Type 1 AGNs is 25%, while it drops to 6% for Type 2 AGNs. The rest of the X-ray selected AGN population presents on average a larger r.m.s. variability than the bulk of non variable sources, indicating that variability detection for at least some of these objects is prevented only by the photometric accuracy of the data. We show how a longer observing baseline would return a larger sample of AGN candidates. Our results allow us to assess the usefulness of this AGN selection technique in view of future wide-field surveys.

Genetic Algorithm Modeling with GPU Parallel Computing Technology

We present a multi-purpose genetic algorithm, designed and implemented with GPGPU / CUDA parallel computing technology. The model was derived from a multi-core CPU serial implementation, named GAME, already scientifically successfully tested and validated on astrophysical massive data classification problems, through a web application resource (DAMEWARE), specialized in data mining based on Machine Learning paradigms. Since genetic algorithms are inherently parallel, the GPGPU computing paradigm has provided an exploit of the internal training features of the model, permitting a strong optimization in terms of processing performances and scalability.

METAPHOR: a machine-learning-based method for the probability density estimation of photometric redshifts

A variety of fundamental astrophysical science topics require the determination of very accurate photometric redshifts (photo-z). A wide plethora of methods have been developed, based either on template models fitting or on empirical explorations of the photometric parameter space. Machine-learning-based techniques are not explicitly dependent on the physical priors and able to produce accurate photo-z estimations within the photometric ranges derived from the spectroscopic training set. These estimates, however, are not easy to characterize in terms of a photo-z probability density function (PDF), due to the fact that the analytical relation mapping the photometric parameters on to the redshift space is virtually unknown. We present METAPHOR (Machine-learning Estimation Tool for Accurate PHOtometric Redshifts), a method designed to provide a reliable PDF of the error distribution for empirical techniques. 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. We present a summary of results on SDSS-DR9 galaxy data, used also to perform a direct comparison with PDFs obtained by the Le Phare spectral energy distribution template fitting. We show that METAPHOR is capable to estimate the precision and reliability of photometric redshifts obtained with three different self-adaptive techniques, i.e. MLPQNA, Random Forest and the standard K-Nearest Neighbors models.

Astroinformatics, data mining and the future of astronomical research

Astronomy, as many other scientific disciplines, is facing a true data deluge which is bound to change both the praxis and the methodology of every day research work. The emerging field of astroinformatics, while on the one end appears crucial to face the technological challenges, on the other is opening new exciting perspectives for new astronomical discoveries through the implementation of advanced data mining procedures. The complexity of astronomical data and the variety of scientific problems, however, call for innovative algorithms and methods as well as for an extreme usage of ICT technologies.

A cooperative approach among methods for photometric redshifts estimation: an application to KiDS data

Photometric redshifts (photo-z) are fundamental in galaxy surveys to address different topics, from gravitational lensing and dark matter distribution to galaxy evolution. The Kilo Degree Survey (KiDS), i.e. the European Southern Observatory (ESO) public survey on the VLT Survey Telescope (VST), provides the unprecedented opportunity to exploit a large galaxy data set with an exceptional image quality and depth in the optical wavebands. Using a KiDS subset of about 25000 galaxies with measured spectroscopic redshifts, we have derived photo-z using (i) three different empirical methods based on supervised machine learning; (ii) the Bayesian photometric redshift model (or BPZ); and (iii) a classical spectral energy distribution (SED) template fitting procedure (le phare). We confirm that, in the regions of the photometric parameter space properly sampled by the spectroscopic templates, machine learning methods provide better redshift estimates, with a lower scatter and a smaller fraction of outliers. SED fitting techniques, however, provide useful information on the galaxy spectral type, which can be effectively used to constrain systematic errors and to better characterize potential catastrophic outliers. Such classification is then used to specialize the training of regression machine learning models, by demonstrating that a hybrid approach, involving SED fitting and machine learning in a single collaborative framework, can be effectively used to improve the accuracy of photo-z estimates.

Gravitational Lens Images Generated by Cosmic Strings

In this review paper the current understanding on the properties of cosmic strings is shortly outlined with spe- cial emphasis on the observational signatures which can be expected both in the optical, through gravitational lensing, and in the radio, through anisotropies in the cosmic microwave background. The experience gathered during the long term in- vestigation of the former candidate CSL-1 is also shortly summarized.

Constraints on sterile neutrino dark matter from XMM–Newton observations of M33

Using archival XMM–Newton observations of the diffuse and unresolved components’ emission in the inner disc of M33, we exclude the possible contribution from narrow line emission in the energy range 0.5–5u2009keV more intense than 10−6–10−5u2009ergu2009s−1. Under the hypothesis that sterile neutrinos constitute the majority of the dark matter in M33, we use this result in order to put constraints on their parameter space in the 1–10u2009keV mass range.