Giuseppe Riccio

Feature Selection Based on Machine Learning in MRIs for Hippocampal Segmentation

Neurodegenerative diseases are frequently associated with structural changes in the brain. Magnetic resonance imaging (MRI) scans can show these variations and therefore can be used as a supportive feature for a number of neurodegenerative diseases. The hippocampus has been known to be a biomarker for Alzheimer disease and other neurological and psychiatric diseases. However, it requires accurate, robust, and reproducible delineation of hippocampal structures. Fully automatic methods are usually the voxel based approach; for each voxel a number of local features were calculated. In this paper, we compared four different techniques for feature selection from a set of 315 features extracted for each voxel: (i) filter method based on the Kolmogorov-Smirnov test; two wrapper methods, respectively, (ii) sequential forward selection and (iii) sequential backward elimination; and (iv) embedded method based on the Random Forest Classifier on a set of 10 T1-weighted brain MRIs and tested on an independent set of 25 subjects. The resulting segmentations were compared with manual reference labelling. By using only 23 feature for each voxel (sequential backward elimination) we obtained comparable state-of-the-art performances with respect to the standard tool FreeSurfer.

An analysis of feature relevance in the classification of astronomical transients with machine learning methods

The exploitation of present and future synoptic (multiband and multi-epoch) surveys requires an extensive use of automatic methods for data processing and data interpretation. In this work, using data extracted from the Catalina Real Time Transient Survey (CRTS), we investigate the classification performance of some well tested methods: Random Forest, MultiLayer Perceptron with Quasi Newton Algorithm and K-Nearest Neighbours, paying special attention to the feature selection phase. In order to do so, several classification experiments were performed. Namely: identification of cataclysmic variables, separation between galactic and extragalactic objects and identification of supernovae.

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.

C 3, A Command-line Catalog Cross-match Tool for Large Astrophysical Catalogs

Modern Astrophysics is based on multi-wavelength data organized into large and heterogeneous catalogs. Hence, the need for efficient, reliable and scalable catalog cross-matching methods plays a crucial role in the era of the petabyte scale. Furthermore, multi-band data have often very different angular resolution, requiring the highest generality of cross-matching features, mainly in terms of region shape and resolution. In this work we present C 3 (Command-line Catalog Cross-match), a multi-platform application designed to efficiently cross-match massive catalogs. It is based on a multi-core parallel processing paradigm and conceived to be executed as a stand-alone command-line process or integrated within any generic data reduction/analysis pipeline, providing the maximum flexibility to the end-user, in terms of portability, parameter configuration, catalog formats, angular resolution, region shapes, coordinate units and cross-matching types. Using real data, extracted from public surveys, we discuss the cross-matching capabilities and computing time efficiency also through a direct comparison with some publicly available tools, chosen among the most used within the community, and representative of different interface paradigms. We verified that the C 3 tool has excellent capabilities to perform an efficient and reliable cross-matching between large data sets. Although the elliptical cross-match and the parametric handling of angular orientation and offset are known concepts in the astrophysical context, their availability in the presented command-line tool makes C 3 competitive in the context of public astronomical tools.

Machine Learning Based Data Mining for Milky Way Filamentary Structures Reconstruction

We present an innovative method called FilExSeC (Filaments Extraction, Selection and Classification), a data mining tool developed to investigate the possibility to refine and optimize the shape reconstruction of filamentary structures detected with a consolidated method based on the flux derivative analysis, through the column-density maps computed from Herschel infrared Galactic Plane Survey (Hi-GAL) observations of the Galactic plane. The present methodology is based on a feature extraction module followed by a machine learning model (Random Forest) dedicated to select features and to classify the pixels of the input images. From tests on both simulations and real observations the method appears reliable and robust with respect to the variability of shape and distribution of filaments. In the cases of highly defined filament structures, the presented method is able to bridge the gaps among the detected fragments, thus improving their shape reconstruction. From a preliminary a posteriori analysis of derived filament physical parameters, the method appears potentially able to add a sufficient contribution to complete and refine the filament reconstruction.

Milky way analysis through a Science Gateway: workflows and resource monitoring

This paper presents the latest developments on the VIALACTEA Science Gateway in the context of the FP7 VIALACTEA project. This science gateway operates as a central workbench for the VIALACTEA community in order to allow astronomers to process the new-generation (from Infrared to Radio) surveys of the Galactic Plane to build and deliver a quantitative 3D model of our Milky Way Galaxy. The final model will be used as a template for external galaxies to study star formation across the cosmic time. The adopted AGILE software development process allowed to fulfill the community needs in terms of required workflows and underlying resources monitoring. The scientific requirements arose during the process highlighted the needs for easy parameter setting, fully embarrassingly parallel computations and large-scale input dataset processing. Therefore the science gateway based on the WS-PGRADE/gUSE framework has been able to fulfill the requirements mainly exploiting the parameter sweep paradigm and parallel jobs execution of the workflow management system. Moving from the development to the production environment an efficient resource monitoring system has been implemented to easily analyse and debug sources of failure due to workflows computations. The results of the resource monitoring system are exploitable not only for IT experts administrators and workflow developers but also for the final users of the gateway. The affiliation to the STARnet Gateway Federation ensures the sustainability of the presented products after the end of the project, allowing the usage of VIALACTEA Science Gateway to all the stakeholders and not only to the community members. Keywords—Workflow Systems; Science Gateways; Collaborative Environments; Astrophysics; DCIs; Milky Way Analysis; Infrastructure Tests; Monitoring

Vialactea Visual Analytics Tool for Star Formation Studies of the Galactic Plane

We present a visual analytics tool, based on the VisIVO suite, to exploit a combination of all new-generation surveys of the Galactic Plane to study the star formation process of the Milky Way. The tool has been developed within the VIALACTEA project, founded by the 7th Framework Programme of the European Union, that creates a common forum for the major new-generation surveys of the Milky Way Galactic Plane from the near infrared to the radio, both in thermal continuum and molecular lines. Massive volumes of data are produced by space missions and ground-based facilities and the ability to collect and store them is increasing at a higher pace than the ability to analyze them. This gap leads to new challenges in the analysis pipeline to discover information contained in the data. Visual analytics focuses on handling these massive, heterogeneous, and dynamic volumes of information accessing the data previously processed by data mining algorithms and advanced analysis techniques with highly interactive visual interfaces offering scientists the opportunity for in-depth understanding of massive, noisy, and high-dimensional data.

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.

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.

C3: 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