The gradual shift in focus from e-infrastructures towards their efficient deplorability leads to new ways of doing science: problem driven, multi-disciplinary, holistic in its deployment of e-infrastructures and software/tools, such as the latest fruits from Computer Science, and – in this particular time frame – often data intensive. This shift in focus and this new way of doing science is embodied in the term escience and in the organizations in support of this development.

The Netherlands eScience Center is solely devoted to this development, but there are many other institutes and centers in The Netherlands, Europe and elsewhere that face the same transition and are in a phase of delivering support, developing services, new tools, new educational directions to enable scientists to accelerate discovery. Therefore new social network platforms are being set up (ePLAN and PLAN-E) to share experiences, best practices, career challenges, curriculum innovation and more.

One of the shared topics is "data stewardship and software sustainability" for which a conceptual approach is being designed, involving researchers at disciplinary level.

back

The life-science and health care research environments offer an abundance of new opportunities for improvement their efficiency and productivity using big data and analytics in collaborative research processes. This talks introduces research results we achieved in two recent research projects, ABA and SPES, respectively.
1) Dataspace-Based Support Platform for Breath Gas Analysis (ABA). We have addressed the development and utilization of an information infrastructure called the ABA-Cloud [1] based on the novel scientific dataspace paradigm, which has aimed at providing advanced data management, automated data collection, data analysis and experimentation within the domain of breath gas analysis. The breath gas analysis scientific community is continuously developing new analytical methods and collecting pilot data, for instance to identify marker compounds for various diseases, like lung cancer, negative side effects of chemotherapies, etc.
2) Support Patients through E-Service Solutions (SPES). Based on the results of the ABA project, we focused on the development of the indoor and outdoor support for dementia patients, non-invasive methods of their brain stimulation, capturing and analysis of the produced data and querying the analysis results [2]. The talk will also discuss the extension of the developed technology to other research domains.

[1] I. Elsayed, et al. ABA-Cloud: Support for Collaborative Breath Research. Journal of Breath Research, 7 (2013) 026007 (13 pp). Online at stacks.iop.org/JBR/7/026007.
[2] P. Brezany, et al. Optimized Management of BIG Data Produced in Brain Disorder Rehabilitation. To appear in the book "BIG Data Optimization", Springer, March 2016.

back

We shall discuss two ambitious projects in computational science which are currently underway, making use of Prometheus and PL-Grid resources. One of these projects is concerned with a grand challenge in materials science — namely, the theoretical prediction of large scale materials properties based directly on the chemical composition of their ingredients. We describe our virtual laboratory for multiscale modelling and simulation in distributed high performance computing environments which makes this possible, through the use of sophisticated software services which handle a very demanding workflow.

The second project is concerned with personalised medicine. The potential impact centres around our ability to marshall substantial quantities of patient data and to use them to perform predictive modelling and simulation in order to deliver therapies and to enhance clinical decision making, on time scales which are far shorter than those usually considered in the context of academic research. Secure access to personal data, as well as to powerful computational resources, are essential.

We shall discuss the underlying e-infrastructure requirements, including data, compute and networks, and describe how we propose to use Prometheus & PL-Grid to support this work.

back

Nowadays, research is increasingly and in many cases exclusively data driven. Knowledge of how to use tools to manipulate research data, and the availability of e-infrastructures to support them, are foundational.

New communities of practice are forming around interests in digital tools, computing facilities and data repositories. By making infrastructure services, community engagement and training inseparable, existing communities can be empowered by new ways of doing research, and can grow around interdisciplinary data that can be easily accessed and manipulated via innovative tools as never done before.

Key enablers of this process are openness, participation, collaboration, sharing and reuse, and the definition of standards and the enforcement of their adoption within the community. In some disciplinary areas communities succeeded in defining and adopting a community governance of the production and sustainable sharing of data and applications.

The European Commission has identified the completion of the Digital Single Market (DSM) as one of its 10 political priorities and much effort is being done in aligning strategies and initiatives along the DSM targets.

This presentation provides an overview of the services and the initiatives that EGI is supporting to promote open science and the technical roadmap and challenges that will drive its future evolution.

back

The mission statement of RDA is: Researchers and innovators openly sharing data across technologies, disciplines, and countries to address the grand challenges of society. The Research Data Alliance aims to building the social and technical bridges that enable open sharing of data and their preservation.  The talk will review the progress so far and discuss current challenges and future plans.

back

Tensor network represents tensors (data cubes) and multi-linear relations between them. In quantum information tensors form a natural way of representing quantum states and operations. Multi-linear relations between tensors provide an extension of matrix analysis, that simplifies greatly reasoning about quantum objects. Tensor networks have found use in many applications in classical informatics. Those networks can efficiently describe large data and relations between them. The talk will cover elements of tensor networks with applications to quantum informatics and data compression.

back

Environmental computing is a relatively novel term describing an integrated approach to producing a comprehensive view of different environmental processes and their interactions and impact. On technical level this means linking tools and approaches from several disciplines together to create a multi-model, multi-data ensembles that can be used in solving different planning, policy or operational challenges. However, in order to effectively support the strategic or operational action, it is important to pay close attention to standards and processes surrounding the technical solution. For example, consistent metadata describing the capabilities and limitations of the models is a crucial feature of environmental computing systems. Similarly, consistent data curation practices and sustainability of the archive services across organizations is similarly essential. The multi-model approach also inherently removes the original researchers from direct contact with the execution of the simulations and analysis tasks, and makes metadata and other supporting material crucial in linking the research and policy/operational activities.
Technical interoperability and consistency of the procedures are further complicated by the environmental computing being a distributed system: data sources (especially sensors) require distributed approach, and amounts of data in different repositories make transferring all the data to single location impossible.

back

The Virtual Physiological Human is an initiative that was supported under the Framework 7 programme, with the aim of developing sophisticated and integrated computational models of human systems and of reaching impact in the industrial and clinical environments. In this talk I will give examples of the types of models and work flows that have been developed and outline their current translational status. I will review the requirement for hpc, grid and cloud computing in the context of exploitation of VPH technology. In the industrial context of device design there is a clear requirement for operation of the most sophisticated models, with associated hpc demands. For the most efficient and effective analysis in a clinical environment there is a great potential for the use of Reduced Order Models, which will often be underpinned by large-scale off-line computing. I will outline some of the remaining challenges, which include the representation of variation and uncertainty in VPH models, and the integration of a wide range of heterogeneous data from clinical and other sources to inform the physiological envelope over which the models should operate.

back

In this talk I will describe an approach to the modelling and simulation of human crowds. By developing accurate models of human crowds the hope is to be able to predict and perhaps manage mass gatherings of people, or enable safer architectural design. There have been numerous attempts at modelling crowds over the past two decades, here I will highlight one approach which uses agent-based modelling. While agent-based modelling offers a very natural approach for modelling human crowds it does introduce significant challenges. Firstly, how to build a model capable of capturing human-like decisions. Secondly, how to calibrate models against real-world data and finally how to validate the model output. This talk will cover work in these areas and finally discuss ways in which these findings might be used to help make large human gatherings safer.

back

During the last 25 years HPC has been considered as a fixed asset. Companies and organizations running HPC workloads gradually realize that moving their HPC infrastructure and operations from a fixed asset to an operating expense model has several advantages.  Despite the technical constraints imposed by the nature of the HPC workloads, there is a business model that today provides companies and organizations a way to move a part or all of their HPC infrastructure to the cloud.  Large European HPC users are currently moving their HPC workloads to the cloud, whatever the cloud implementation could be. In cooperation with Intel HP has a capability to provide a variety of solutions in this space, with adapted technical and financial models. We will provide some examples of companies which have made this move, and will also provide some hints and guidelines to support this model.

back

Since its re-launch in 2010, Software Carpentry has become a world-wide organization: between January 2013 and July 2015, its volunteer instructors delivered over 400 workshops to over 12,000 researchers in two dozen countries. Along the way, it has developed an instructor training program that introduces scientists to evidence-based teaching practices, and shown that development practices originating in open source software can successfully be applied to the development of educational materials. This talk will explore the lessons we have learned along the way and discuss some of the experiments we would like to try next.

back

Big data have attracted great attention in science, industry and society. However, its nature and the fundamental challenge have not been recognised. Big data research is developing with the evolving scientific paradigm, the fourth industrial revolution, and transformational innovation of technologies. This lecture tries to answer the following questions: What is the nature of big data? What kind of infrastructure is requested to support not only big data management and analysis but also knowledge discovery and management? What is the fundamental challenge of big data computing? Can we find an approach to mapping big data into knowledge space?

back