The Challenge of Big Data in Science (5th International LSDMA Symposium)

Wednesday 5 Oct 2016, 08:30 → 17:00 Europe/Berlin

Aula, FTU (KIT)

Management of scientific Big Data poses challenges at all stages of the data life cycle – acquisition, ingest, access, replication, preservation, etc. For scientific communities the data exploration – commonly considered as the 4th pillar besides experiment, theory and simulation – is of utmost importance to gain new scientific insights and knowledge.

At this symposium, organized by the cross-program initiative “Large-Scale Data Management and Analysis” (LSDMA) of the German Helmholtz Association, international experts address several aspects of Big Data. It also provides a common platform for discussions and identification of new perspectives.

The participation fee of 50 EUR is to be paid in cash at arrival (for KIT-CN employees an alternative payment method will be offered) and includes coffee, lunch and shuttle bus service. For further info, please contact lsdma@scc.kit.edu

08:30 → 09:00 Registration

09:00 → 09:30 Welcome, Introduction

Speaker: Prof. Achim Streit (KIT)

Slides 20161005_LSDMA_Grußwort.pptx LSDMA-Symposium-Opening.pptx

09:30 → 10:30 Moving data from research enabler to trusted research output: challenges for infrastructure providers and national services

For most of the last 200 years, data have been seen as an adjunct to the process of publishing research outputs (articles, conference papers, etc). The data were there to inform the creation of new knowledge, and to enable the production of tables and graphs to appear in publications. Over the last 10 years, data have steadily moved towards becoming first class research outputs. But what does this development mean for research infrastructure and national service providers? The talk will argue that research infrastructure has fundamentally always been about data, and that the role of data in the scholarly communication ecosystem can be seen in terms of a modified set of the fundamental functions required by such a system. The talk will then consider the Australian National Data Service as case study of a national response to these issues, before concluding by talking about the importance of trust both in research outputs and in the locations where these outputs are made available.

Speaker: Andrew Treloar (Australian National Data Service)

Slides Treylor,_LSDMA_Keynote,_October_5_2016.pptx

10:30 → 11:00 Coffee break

11:00 → 11:30 Funding Research Data Management and related Infrastructures

Funding and policy formulation among Research Funding Organisations (RFOs) and Research Performing Organisations (RPOs) with respect to Research Data Management and Access to Research Data is an ongoing concern. It is taking place at different paces among countries and organisations, and is part of funding RD related infrastructures. In 2016 Knowledge Exchange and Science Europe published a briefing paper based on analysis of a survey on current practice and views in European RFOs and RPOs – main questions: where does the funding come from, how is the funding organised, what are the challenges. Results, conclusions and recommendations will be presented.

Speaker: Bas Cordewener (Knowledge Exchange)

Slides KE-SE_Funding_RDM_and_RDI_Karlsruhe_05102016_v2.pptx

11:30 → 12:00 Getting Science out of eScience

Research questions in modern research are increasingly addressed using data and compute intensive methodologies. eScience bridges the scientific ambitions of researchers and modern digital tools and knowledge from data and computer science. I will take two approaches to illustrate how eScience functions. First, I will show applications from different scientific domains where Big Data and eScience methodologies are used. The value of eScience is shown by accelerating pulsar search pipelines in astronomy, data assimilation to predict weather at the scale of streets in Amsterdam, visualizing large point clouds and text analysis in historical documents. Second, I will show that eScience itself evolves towards a scientific domain. The research question aware development of methodologies leads to a cycle of tailoring technologies to specific research questions and generalizing them again. At the Netherlands eScience Center the eScience Technology Platform (eStep) has been set up to bring the knowledge, best practices and the generic software together.

Speaker: Wilco Hazeleger (Netherlands eScience Center)

Slides https://www.youtube.com/watch?v=x1SgmFa0r04 NLeSC_4KIT.pptx

12:00 → 12:30 The role of the EDISON Framework in building Data Science professionals: one size definitely does not fit all

The emergence of Data Science technologies is having an impact on nearly every aspect of how research is conducted, how scientists think, and how research data are used and shared. The EU-funded EDISON Project is putting in place foundation mechanisms that will speed-up the increase in the number of competent and qualified Data Scientists across Europe and beyond. The EDISON initiative is doing this through various measures aimed at reducing the gap between the supply side of educators and trainers and the demand side of employers. The EDISON Data Science Framework (EDSF) is based around a group of tightly related documents namely the Competence Framework, the Body of Knowledge and the Model Curriculum; the various processes that exist to keep the information within these documents fresh and relevant are almost equally as important.

Speaker: Steve Brewer (University of Southampton)

Slides LSDMA-Brewer-5-10-2016.pptx

12:30 → 14:00 Lunch break

14:00 → 14:30 The development of national road maps for digital research infrastructures in the international context

Since the introduction of Hollerith machines there have been ideas for the evaluation of time wise massive data collections. However, the main developments started in the sixties based on main-frame computing. Already at that time many dreams were published which took decades to come through. As the delivery of promises and hopes was much less success-ful than expected many early starters and their institutions were shut down. Driving coun-tries in this competitive game were not only the US and other industrialized countries but also highly funded national institution in many other countries. The complexity of the tech-nology and the dynamics of software development forced most developments out of the competitive game. The US approaches became very dominant due to several "cultural" fac-tors like massive funding ressources and well funded disruptive approaches. However, the development of digital research infrastructures in different scientific disciplines has been driven by very specific national factors, of which available IT-power is only one of many im-pact factors. Intenational research competition is more and more based on the management of digital research infrastructures. Easy answers and simple explanations of how to move ahead are not available.

Speaker: Otto Rienhoff (Uni Göttingen)

Slides KIT-2016-big-data-Rienhoff-fin.pdf

14:30 → 15:00 User Community Driven Development in Trust and Identity

During the last decade, national Identity Federations for Research and Education have emerged globally. There is a growing interest from different research communities in using federated access, although some challenges still exists that prevent the wide usage of this approach. Federation operators have invested significant efforts to enhance federated technologies and policies to support community requirements; and whilst some requirements get implemented, new requirements emerge. The AARC EC-funded project which started in May 2015, aims to define building blocks and policy best practices to enable users to seamlessly access research and e-infrastructure services and to support new research collaboration to build interoperable AAIs. This talks aims to provide an overview on the user-driven developments in Europe and how AARC is addressing them. Obviously research and collaboration is global; it is therefore essential to think global (even if funding is local) to ensure that relevant activities are synchronized internationally.

Speaker: Christos Kanellopoulos (GRNET)

Slides 20161005-LSDMA-AARC-User-Community-Driven-Development-in-TI.pptx

15:00 → 15:30 Coffee break

15:30 → 16:00 Assessing Societal Aspects of Big Data - the Project ABIDA

The interdisciplinary project ABIDA (Assessing Big Data), funded by the Federal Ministry of Education and Research, explores social opportunities and risks of the generation, linking and analysis of huge amounts of data and develops options for political action, research and development. ABIDA approaches the topic of Big Data from a fundamentally interdisciplinary perspective. Sociologists, philosophers, economists, legal and political scientists work hand in hand on this. The project aims to jointly gather existing knowledge about dealing with big data, generate new knowledge, deepen the knowledge, and make it accessible to the public. The scientists will examine the societal impacts associated with Big Data by using the methods of technology assessment oriented to dialogue and participation.

Speaker: Reinhard Heil (KIT)

Slides Assessing_Societal_Aspects_of_Big_Data_-_the_Project_ABIDA.pptx

16:00 → 17:00 Beyond the "under-desk datacentre": lessons from research data management in the UK

In July, four of the UK leading research organisations (HEFCE, Research Councils UK, Universities UK and the Welcome Trust) launched a concordat on open research data (http://www.rcuk.ac.uk/media/news/160728/), outlining ten principles including: the importance of developing data skills; the importance of ensuring data underlying publications is accessible by publication date; the rights of data creators to reasonable first use; and the expectations of data users to acknowledge use of others’ data. This is a significant step in a long journey in both acknowledging the importance of sharing research data in research policy and practice, as well as the providing the tools and services needed to make this possible. Jisc has been working with research organisations and universities in the UK both in advocating open research data, and in creating institutional and national tools and services within the context of international research e-infrastructures. Currently Jisc is working with UK Universities to develop a research data shared service. This talk will outline the work Jisc has undertaken to support research data management in the UK and to encourage and support researchers not to lock data away on USB sticks or on a PC under the desk, and the lessons learnt along the way.

Speaker: Matthew Dovey (JISC)

Slides lsdma-2016_Jisc.pptx