I-science workshop on data mining, distributed computing and visualization for astronomy

Announcement

I-science for astronomy

I-science workshop on data mining, distributed

computing and visualization for astronomy.

Aimed at advanced master students, graduate students and postdocs to stimulate

innovative collaborative computer science developments. Staff is also very welcome.

Three personal NWO research grants will be granted at PhD/postdoc level following

the workshop.

Application deadline: 15 September 2008

via http://www.lorentzcenter.nl/lc/web/2008/321/application.php3?wsid=321

Lorentz Center, Leiden, The Netherlands

week 13 Oct – 17 Oct 2008

Keywords

Visualizing the universe

Streaming processing

Computing the universe

The astronomical virtual observatory

Aim of the workshop

The I-science for astronomy workshop will stimulate the interaction between computer

scientists and astronomers by exploring innovative astronomical applications for data mining,

data modelling, distributed computing and visualization (for more information about I-science

funding programme, please see the appendix). The workshop objectives are to highlight the

role of computer science in astronomical research, including advanced data mining, powerful

distributed computing, data visualization and the virtual observatory. This I-science initiative

will bring together astronomers and computer scientists in a collaboration programme that

starts with a one week workshop. During the workshop new innovative collaborative research

projects will be explored and perceived; a selection of these will continue as NWO funded

scientific research projects, either for postdocs or for PhD projects (total budget 0.5 M€). An

independent committee will judge the quality of proposals after the workshop.

The workshop aims to enhance collaborations between astronomy and computer science by

nurturing cross disciplinary initiatives, with a particular focus on motivating students, PhDs

and postdocs for innovative I-science developments. Astronomy offers big challenges to

computer science, which require fundamentally new methods or even abstractions.

The originality of this workshop is marked by the following:

-        addressing cutting edge computer science research issues;

-        delivering a number of research project proposals and initiatives taking the existing I-

      science projects as a source of inspiration;

-        involving a truly interdisciplinary cooperation (involving three I-science research

      programmes: GLANCE, VIEW and STARE research programme);

-        involving  both advanced master and PhD students as well as post-doc and senior

      researchers;

-        hosting world leading keynote speakers in the field.

Organizers

Organizers: Edwin Valentijn (RUG, astronomy, astronomical information technology), Jack

van Wijk (TU/e, computer science, data visualization), Farhad Arbab (CWI, computer

science, distributed computing), Cees de Laat (UvA, computer science, distributed

computing).

Co-organizer: Jos Roerdink (RUG, computer science, data visualization)

Participation

The workshop should attract master students, advanced PhD student and postdocs. Staff is

also very welcome.

Application deadline is 15 September 2008. A short statement of interest and CV should be

filled in at the Lorentz center website:

http://www.lorentzcenter.nl/lc/web/2008/321/application.php3?wsid=321.

Pre-selection of the applicants will take place in the week after this deadline, ensuring that

the they meet a minimum requirement. On 22 September 2008 the applicants will be

informed whether their request for participation has been approved or not. Invited

participants will thereafter have to fill in the registration form at the Lorentz Center website.

For all invited participants, the Lorentz Center offers to pay accommodation and travel

expenses.

Grants for participants

All participants are eligible for travel and housing grants.

Research grants

The participants who want to compete for the research grants will design and write research

proposals immediately after the workshop. The non-candidate members of the team support,

elaborate, and nurture the ideas of the candidate; they are not exclusively connected to only

one team and supporting only one candidate. Three full days after the workshop, candidates

will submit one-page project pre-proposals. Pre-proposals will be quickly selected and the

most promising pre-proposals will be nurtured into full proposals. Participants asked to

present a full proposal, will be offered a workshop at 4 November 2008. They will compete

for three personal PhD/post-doc grants. For the three grants budget of 0.5 M€ is available.

Location

This workshop will take place at the Lorentz Center, an international visitors’ center of Leiden

University. We conduct workshops in the fields of physics, mathematics, astronomy,

computer science and life sciences.

During their stay all participants will have the use of a desk and a computer account in one of

the 20 offices at the Lorentz Center. All offices are equipped with a whiteboard, desks and a

PC running Linux and Windows, and access to the wireless Internet. In addition, each office

has cable connections for a laptop to the Internet. The Center has several lecture rooms, a

meeting room and an informal ‘common room’ for having a cup of coffee or tea, for

discussing, chatting or reading.

Science libraries, cafeteria, photocopiers and other such facilities are nearby, as the center is

fully integrated within the buildings of Leiden University. For more information, please, visit

our website: www.lorentzcenter.nl

Description of the workshop

During the workshop, half of the time will be devoted to presentations by keynote speakers

on issues on the border of astronomy and computer science. Speakers will explore the open

issues and reach out for the definition of innovative research projects. Presentations will be

given by astronomers and computer scientists from the Netherlands and abroad. The other

half of the time, the workshop participants will team up into small groups working together

identifying the nature, importance and true bottlenecks of the presented problems. Speakers

will define some starting point or exercises for this and are expected to optionally participate

in the teams. The fraction of time reserved for the small working groups will gradually

increase during the week.

During the workshop the programme will connect and integrate the seemingly dissociated

fields: visualization of astronomical data, streaming and grid processing of astronomical data,

database modelling of astronomical data, fundamental physical limits and the information

theory such as encountered in the astronomical setting and the astronomical virtual

observatory in action. This is an ambitious programme indeed, meant to provide the source

of inspiration for fundamental cross bordering astronomical information theory research.

Other example issues are:

-         visualization of multidimensional parameter space of very large data sets.

Current information technology development coupled to modern astronomical instrumentation is leading to accelerated acquisition of large volumes of observational

data stored in online archives. They contain high quality images and multi-dimensional

parameter values of astronomical objects. Each of these provides catalogues of many

millions of objects, with often several tens of parameters characterizing the objects. We

need to develop new approaches for analyzing and visualizing astronomically interesting information out of the flood of data. No adequate tools are yet available in the astronomy

research community for extracting and visualizing multi-dimensional data. The need for

such tools is especially crucial in the era of very large data bases (VLDB's) with tens of

Tbytes of data on millions of objects with multi-dimensional data. The necessary tools

could become available when full use is made of current knowledge in the areas of (nonlinear) image processing, pattern recognition and visualization

-         pattern recognitions using wavelets in astronomical images

-         extreme data modelling, using linking and backward chaining leading to a much more

            advanced reference systems than Google

-         set the conditions to realize a data base environment allowing changing computational

methods, such as astronomers really  do in practice

-         sensor networks like Lofar deliver data rates which require innovative approaches for

      streaming, processing and networking data.

-         real time distributed correlation and the application in e-VLBI. Seeking methods to

distribute massive dataflow over a grid of compute nodes, in order to calculate the

astronomical correlation functions between all telescope elements. The challenge is to

level the resources in such way that the(non-uniform) computer resources can keep up with

the time-critical data flow.

-         what are the fundamental physical limits in astronomical computing - is the universe computing,

      can we make use of that?

-         using the astronomical  Virtual Observatory to do dynamic research operations (i.e

      deriving results by computing), work flows etc

-         image reconstruction (point spread function homogenization) for optical+infrared data for

      deriving very accurate photometric redshifts of billions of galaxies

-         eVLBI In eVLBI radio telescope signals are combined (correlated) to mimic a much larger

instrument. Nowadays routinely signals from telescopes across Europe and USA are

brought together and processed in a central facility. In the old days tapes and discs were

used, nowadays Gigabit and faster networks transport the data. Next step is to move

from a fixed dedicated special purpose hardware computing engine named the correlator

to a more generic grid infrastructure. This brings new problems in bringing network and

distributed computing resources together to process the huge constant stream of

information. In this session speakers will outline the current state of this work and present

the challenges imposed on the networks and compute clusters that need to be solved.

-         Cosmological model simulation on supercomputers across the globe coupled by high

speed networks. For the simulation of a significant "piece of universe" several supercomputers will be

coupled together such that each can process its own piece while exchanging boundary

information with the other computers in the set. At current processor speeds the

overhead in networking due to round trip time are significant if not handled correctly and

treated in the overall computation model. This session will outline the model calculations,

the segmentation of the problem space and the problems that still need be solved.

Appendix: The NWO I-science funding programme

The workshop is organized by scientific leaders of the e-science projects being carried out

under the NWO programme I-science. I-science is a NWO cluster of three research

programmes (GLANCE, VIEW and STARE) that started in 2005 and will end in 2011. I-

science is focussed on empowering of the computer science discipline in the Netherlands

and creating of synergy between involved research groups. The programme cultivates the

research in the field of e-science aiming to bringing up the Netherlands e-science to a

prominent position. The I-science research cluster is tackling this goal by funding at the first

place 17 relatively big research projects in the filed of the large scale distributed and parallel

systems (GLANCE), visualization (VIEW) and collaborative projects between astronomy and

computer science (STARE). But besides the projects itself I-science cluster is also facilitating

communication activities between the projects in various ways – by means of conferences

and workshops like this one. The I-science programme text can be downloaded from the

following link:

http://www.nwo.nl/files.nsf/pages/NWOP_6H3CD4/$file/I-Science%20totaal.pdf

Contact info

Prof. dr Edwin A. Valentijn (RUG, astronomy)

KAPTEYN INSTITUTE P.O. Box 800 NL-9700 AV Groningen, The Netherlands telephone: +31 (0)50 3634011 mobile: +31 (0) 6 48276416

valentyn@astro.rug.nl

Prof. dr Jack van Wijk (TU/e, computer science)

Department of Mathematics and Computing Science Eindhoven University of Technology P.O. Box 513 5600 MB Eindhoven, The Netherlands

telephone : +31 40 247 4579 telephone :  +31 40 247 5010 (Secr.)

vanwijk@win.tue.nl

Prof. dr Farhad Arbab (CWI, computer science)

Kruislaan 413 1098 SJ Amsterdam, The Netherlands telephone: +31-20-592-4056 telephone: +31-20-592-4189 (Secr.)

Farhad.Arbab@cwi.nl

Dr Cees de Laat (UvA, computer science)

Faculty of Science, Informatics Institute, University of Amsterdam,  Kruislaan 403, room F241 NL-1098 SJ Amsterdam, The Netherlands telephone : +31 20 525 7590 mobile : +31 6 5156 6438 delaat@uva.nl

Prof. dr Jos Roerdink (RUG, computer science)

University of Groningen Institute for Mathematics and Computing Science P.O. Box 800, 9700 AV Groningen, The Netherlands telephone: +31 (0)50 363 3931 telephone: +31 50 363 3939 (Secr.) j.b.t.m.roerdink@rug.nl