Início » Tecnologia & Inovação » Webinars Técnico » 2018
08/11 - 02:00 pm
Jordan Raddick
Abstract ⓘ
Abstract
×Speaker
Jordan Raddick
Title
Title: A new universe of online data (The Johns Hopkins University)
Abstract
SciServer (www.sciserver.org) is an online system for scientific research and education with big data. The system offers access to several Petabyte-scale scientific datasets in fields ranging from astronomy to turbulence to genomics, along with a set of simple but powerful browser-based tools to visualize and analyze those datasets. The power of SciServer lies in its close connection to datasets like the Sloan Digital Sky Survey (SDSS), allowing for simple, fast analysis with minimal setup. The SciServer Dashboard offers quick access to all SciServer tools and datasets, and to all prior work within the system. All users up to 10 GB of file and database storage space – and through the Dashboard, users can manage uploaded data files, and can create collaborative groups to share datasets and scripts publicly or with selected colleagues. SciServer makes it easy to share exactly the right data with exactly the right people. SciServer’s Compute tool enables online computational analysis of big data through Jupyter notebooks. With only a few clicks, a user can create a customized environment on fast, high-memory virtual machines hosted by IDIES, where they can write or upload Python, R, or Matlab scripts to perform all sorts of data-intensive tasks. Importantly, SciServer allows these scripts to be run in either interactive or batch mode, meaning that even the most demanding data-intensive analyses can be completed with ease. To help scale up research quickly, SciServer Compute features a wealth of “getting started” data files and example notebooks that can be adapted to meet researchers’ needs. In addition to SciServer’s obvious potential for data-intensive science and engineering research, the system has been used effectively in many educational settings, particularly as data-intensive lab activities for undergraduate science courses. Learners get immediate access to high-performance computing resources without the need to install and configure software. A new set of “SciServer Courseware” notebooks creates an environment to manage learning activities by creating shareable user volumes and groups so that students and TAs can access files with appropriate permissions. SciServer is funded by National Science Foundation award ACI-1261715. For more information, please visit www.sciserver.org or contact the SciServer Helpdesk at sciserver-helpdesk@jhu.edu
Slides not available
25/10 - 01:00 pm
Keith Bechtol
Abstract ⓘ
Abstract
×Speaker
Keith Bechtol
Title
LSST Commissioning Science Verification and Validation (LSST)
Abstract
As the integrated Large Synoptic Survey Telescope (LSST) system comes together during commissioning, we need to understand the degree to which the camera, telescope, and data management system are functioning together in a way that will support the high-level scientific goals of the 10-year survey. Accordingly, a major goal of commissioning is to quantify the distribution of performance of the as-built LSST system using a combination of on-sky data, informed simulations, and external datasets. These studies can then be used to inform early operations. I will describe the systems engineering approach adopted by the LSST Project for verification, validation, and characterization, and discuss the current status and plans for sustained data-taking and analysis campaigns during commissioning.
11/10 - 11:00 am
Celso Mendes
Abstract ⓘ
Abstract
×Speaker
Celso Mendes
Title
Current trends in High-Performance Computing (INPE)
Abstract
High-Performance Computing (HPC) is an area where numerous exciting advances have been observed recently. Besides their concrete economic impact to society in general, those advances have truly improved the rate of scientific discoveries. In fact, the use of high-performance machines for simulations has become prevalent in most developments in science and engineering today. In this talk, we will briefly describe the status of existing HPC systems, in Brazil and abroad, pointing to the technological factors that enabled this scenario, including Moore’s Law and multi-core processors. We will then present the major paradigms that are presently used to program these systems, and examples of large scientific applications that have been developed under some of these paradigms. Finally, we will close with a personal view of current trends and challenges that the HPC community will face in the near future, and potential directions for further progress.
25/09 - 01:30 pm
William O\\\'Mullane
Abstract ⓘ
Abstract
×Speaker
William O\\\'Mullane
Title
LSST Data Management Overview and Status (LSST)
Abstract
In this talk a brief update on the status of LSST in general as well as DM specifically will be given. We shall go into some details on the structure and organisation of Data Management.
20/09 - 02:00 pm
Victor Krabbendam
Abstract ⓘ
Abstract
×Speaker
Victor Krabbendam
Title
LSST Project Status: The System is becoming Real (LSST)
Abstract
LSST started construction in 2014 after a decade of consideration, design and early development.There has been tremendous progress in the last 4 years to make the full LSST system a reality. This talk will review the challenging objectives of the LSST and the current progress and achievements across the Project. LSST remains on schedule to begin the wide, fast, deep optical survey of the visible sky starting in October 2022.
13/09 - 11:00 am
Marko Simonovic
Abstract ⓘ
Abstract
×Speaker
Marko Simonovic
Title
Analytical Tools for Future LSS Surveys (CERN)
Abstract
Abstract: I will briefly review the perturbation theory approach to clustering of matter and galaxies and show how well the theoretical predictions match the simulations. I will particularly focus on the so called IR resummation and predictions for the shape of the BAO peak. I will also discuss the limitations of perturbative calculations and how the theoretical uncertainties impact the inference of cosmological parameters in galaxy surveys. Finally, I will mention some of the open questions in analytical modelling of LSS and strategies to resolve them.
23/08 - 11:00 am
Ezequiel Treister
Abstract ⓘ
Abstract
×Speaker
Ezequiel Treister
Title
Chilean Participation in LSST (Pontificia Universidad Catolica de Chile)
Abstract
In this informal talk I will give a brief overview of the LSST project and present an update of the status of its construction. I will then describe the Chilean participation on LSST and our involvement in the different scientific collaborations, with some focus on the AGN Science Collaboration, of which I am a member. Finally, I will present the current data rights and data access plan and how can it impact international contributors, leaving time for questions and comments.
09/08 - 11:00 am
Rafael Santos
Abstract ⓘ
Abstract
×Speaker
Rafael Santos
Title
You\\\'re already a Data Scientist, now go ask for a raise (INPE)
Abstract
There has been a lot of hype around Data Science -- online academic and technical courses, videos, books, blogs, job opportunities, etc. It seems that Data Science is the dream job of the (very near) future, and there are huge opportunities and salaries for data scientists. In this talk I will present different views of what Data Science is and what is expected from a data scientist, trying to filter out the hype. While there is a need for professionals specialized in extracting knowledge from data, we will see that most of the knowledge required to be a data scientist is already being employed by researchers in areas like astronomy, bioinformatics, remote sensing, banking, etc.: in these cases Data Science is more useful as an organizational approach for data-based experiments. This is a non-technical talk, and is the first lecture in a graduate-level course on Data Science at INPE, the Brazilian National Institute for Space Research.
02/08 - 11:00 am
Katrin Heitmann
Abstract ⓘ
Abstract
×Speaker
Katrin Heitmann
Title
Cosmological Simulations for Large-Scale Sky Surveys and the LSST DESC Data Challenge 2 (Argonne National Laboratory)
Abstract
Cosmology -- the study of the origin, evolution, and constituents of the Universe -- is now entering one of its most scientifically exciting phases. Three decades of surveying the sky have culminated in the celebrated \\\"Cosmological Standard Model\\\". Yet, two of its key pillars, dark matter and dark energy -- together accounting for 95% of the mass-energy of the Universe -- remain mysterious. Next-generation observatories will open new routes to understand the true nature of the \\\"Dark Universe\\\". These observations will pose tremendous challenges on many fronts -- from the sheer size of the data that will be collected (more than a hundred Petabytes) to its modeling and interpretation. The interpretation of the data requires sophisticated simulations on the world\\\'s largest supercomputers. The cost of these simulations, the uncertainties in our modeling abilities, and the fact that we have only one Universe that we can observe opposed to carrying out controlled experiments, all come together to create a major test for our future data analysis capabilities. In this talk I will discuss an ambitious end-to-end simulation project that attempts to provide a faithful view of the Universe as seen through the Large Synoptic Survey Telescope (LSST). This simulation project is the foundation for the second data challenge (DC2) carried out by the LSST Dark Energy Science Collaboration (LSST DESC). I will briefly describe how these complex, large-scale simulations will be used in order to prepare the collaboration for the arrival of LSST data.
12/07 - 02:00 pm
Gregory Dubois
Abstract ⓘ
Abstract
×Speaker
Gregory Dubois
Title
The LSST Science Platform - Delivering Data, Supporting Analysis (Caltech)
Abstract
The LSST Science Platform is the set of interfaces that will be provided to the LSST data, user storage, and user computing. The presentation will cover the three “Aspects” of the Science Platform: the data access APIs, providing IVOA-standard and other interfaces to the LSST catalog and image data; the Portal Aspect, providing a Web interface for discovering, querying, and visualizing LSST data products and user data; and the Notebook Aspect, providing a JupyterLab environment for user-driven data access and analysis in Python. I will describe and demonstrate the features already present in the current prototypes and the capabilities still planned for future development.
Slides not available
05/07 - 11:00 am
Matias Carrasco
Abstract ⓘ
Abstract
×Speaker
Matias Carrasco
Title
Astronomical Data Access in the Era of Scientific Cloud Computing (University of Illinois)
Abstract
Data Access is such a simple term to understand but can hide many layers of complex systems and ideas thus it can be unequivocally be interpreted and explained in many different ways in regards software, technology, abstraction or design. This is also true in the context of astronomical data, both literally and figuratively. In this talk I will discuss a personal view on what data access means for astronomical surveys and how we can improve it, on how scientific cloud computing is rapidly evolving and being adopted and on how the revolution current open source technologies like containers, Jupyter and micro services are changing the way we understand and do Astronomy.
07/06 - 02:00 pm
Andrew Connolly
Abstract ⓘ
Abstract
×Speaker
Andrew Connolly
Title
The Rise of the Machines in Astronomy (University of Washington)
Abstract
Machine learning has become integral to the analysis of astronomical data over the last decade. It has aided in a number of ground breaking results; from the identification of early onset supernovae to the efficient selection of QSOs from large survey data sets. In this talk I will describe some of the developments in machine-learning applications that impacted astronomical discoveries, describe why some techniques that, at first, appeared superior to existing applications did not gain traction in the astronomical community, and look at emerging methodologies that may impact astronomical surveys such as the Large Synoptic Survey Telescope
26/04 - 11:00 am
Dominique Boutigny
Abstract ⓘ
Abstract
×Speaker
Dominique Boutigny
Title
The LSST data processing and data analysis centers in France (National Institute of Nuclear and Particle Physics)
Abstract
Together with NCSA, the IN2P3 Computing Center (CC-IN2P3) in France will be processing 50% of the LSST data to produce the annual Data Releases. Both computing centers will then exchange their respective data sets to constitute a complete data archive at both sites. In parallel, IN2P3 is also planning to deploy an LSST Data Access Center with analysis capabilities, targeting the science community. These projects contains several technical challenges that I will describe in details. I will also expose the strategy that we have put in place in order to check that the designed technical solutions are actually matching the scientific needs.
19/04 - 03:00 pm
Melissa Graham
Abstract ⓘ
Abstract
×Speaker
Melissa Graham
Title
LSST Data Products (University of Washington)
Abstract
The LSST will begin preliminary commissioning next year, and is projected to start its 10-year survey of the southern sky in 2022. I will provide a brief overview of the main science drivers and survey design, and give a status update of the project with a focus on the LSST\\\\\\\'s Data Management system: the data processing pipelines, the products they generate (images, catalogs, and alerts), the planned user interface for scientific queries and analysis, and the data access timescales and policies that are relevant to Brazilian astronomers. I will also cover the current activities of the LSST Science Collaborations, which are actively engaging the community to prepare for LSST, and their existing channels for communication and participation. Finally, I will outline the open opportunities for future users to help LSST refine its observational strategies for the main survey (wide-fast-deep), mini-surveys, and deep drilling fields.
12/04 - 01:00 pm
Francisco Javier
Abstract ⓘ
Abstract
×Speaker
Francisco Javier
Title
Introduction to the DESC Data Challenge 2 (University of California, Irvine)
Abstract
With the advent of the next generation galaxy surveys, a need for higher precision modeling and systematics control arises. End-to-end simulations are a great tool to achieve these goals. Though more traditional in particle physics, in the recent years there has been an effort from the astronomical/cosmological community to generate high-fidelity simulated astronomical images. The Dark Energy Science Collaboration (DESC) is now involved in the so-called Data Challenge 2 (DC2), which constitutes a collaboration-wide effort to generate synthetic mocks and simulated images to perform end-to-end analyses. I will give a detailed overview of this data challenge: what it is, what the goals are and how DESC collaborators can get involved.
05/04 - 10:00 am
Andreas Wicenec
Abstract ⓘ
Abstract
×Speaker
Andreas Wicenec
Title
Data Intensive Astronomy Primer (University of Western Australia)
Abstract
In this presentation I will try to bridge historic efforts of humankind to gather and use knowledge and understanding of astronomical phenomena to the modern era of collecting and analysing exponentially growing amounts of digital data while keeping the number of scientists almost constant. I will describe the challenges posed by this obvious bottleneck and also present work we are undertaking in tackling them in order to enable scientists to efficiently extract knowledge.
29/03 - 11:00 am
Kyle Chard
Abstract ⓘ
Abstract
×Speaker
Kyle Chard
Title
Parsl: Parallel Scripting in Python (University of Chicago)
Abstract
Scientists are increasingly reliant on the analysis of large datasets and use of parallel computing resources. While scientific workflows have long been used to simplify the construction of parallel applications, few systems are well equipped to handle new forms of analysis such as those presented by machine learning and data analytics, interactive analysis in Jupyter notebooks, and online computing. The task of developing and executing a variety of analyses at scale can be difficult, requiring complex orchestration and management of applications, data staging across wide area networks, and customization for specific execution environments. In this talk we introduce Parsl (Parallel Scripting Library), a pure Python library for orchestrating the concurrent execution of multiple tasks that is designed to meet the needs of new analysis modalities. Parsl is remarkable for its simplicity. Developers simply annotate functions (either pure Python or wrappers around shell programs) and write application logic to link these functions in Python. Calls to these annotated functions then result in the creation of “tasks” that run concurrently with the main program and other tasks, subject to dataflow constraints defined by the availability of input data. Parsl abstracts hardware details, allowing the same script to be executed efficiently on one or more laptops, clusters, clouds, and/or supercomputers.
16/02 - 11:00 am
Mark G. Allen
Abstract ⓘ
Abstract
×Speaker
Mark G. Allen
Title
CDS - an astronomy data centre for reference data (Strasbourg astronomical Data Center)
Abstract
The CDS is an astronomy data centre that develops services that are widely used by the international community: SIMBAD, the reference database for nomenclature and bibliography of astronomical objects (outside the solar system), VizieR, the service for astronomical catalogues and tables published in journals, the Aladin interactive sky atlas and VO portal, and the CDS X-Match large catalogue cross-correlation service. The content of the CDS is built with strong collaborative relationships with the astronomy journals, observatories, space agencies and projects. The CDS services are compliant with IVOA standards, and the CDS helps build the astronomy data e-Infrastructure via contributions to the definition of the standards. Recent advances in all-sky VO technologies have enabled new innovative ways of using astronomy data, one example being the CDS Aladin-Lite tool and the Hierarchical Progressive Surveys (HiPS) system for the sharing of large survey data. This system is being widely implemented such as in ESASky, JAXA, NOAO and many more, forming a HiPS network and being standardised at IVOA. It is also an integral part of the newly released CDS Portal. In this presentation I will highlight recent advances, the challenges, and future directions of the CDS.
25/01 - 03:00 pm
Stephanie Juneau
Abstract ⓘ
Abstract
×Speaker
Stephanie Juneau
Title
Science & Discovery with the NOAO Data Lab (NOAO)
Abstract
As we keep progressing into an era of increasingly large astronomy datasets, the NOAO Data Lab (datalab.noao.edu), is developing a suite of analysis tools to maximize potential for discoveries and extraction of scientific knowledge. The services are organized around large astronomy surveys – both in-house and external – and include databases, query managers, virtual storage space, workflows through our Jupyter Notebook server, or scripted analysis with the ‘datalab’ command. Working at the intersection of astronomy and data science, our main goal is to provide users with tools to work close to the data, and thus optimize the scientific productivity of the astronomy community. We currently host datasets from NOAO facilities such as the Dark Energy Survey (DES), the DESI imaging Legacy Surveys (LS), the Dark Energy Camera Plane Survey (DECaPS), and the nearly all-sky NOAO Source Catalog (NSC). We are further preparing for large spectroscopic datasets such as DESI. After an overview of the Data Lab and datasets, I will showcase scientific applications on topics such as star/galaxy separation, finding dwarf galaxies, and extragalactic large-scale structures. I will also present steps for new users to get started conducting their own analysis, and describe our vision for future developments as we tackle the next technical and scientific challenges.