The main goal of the “Enabling Computational & Data-Intensive Science and Engineering” topic is to use large, inconsistent scientific data sets, data-driven models, and complex theories to achieve pioneering breakthroughs for science, technology and society.
To do so, Jülich researchers are working in close collaboration with the Karlsruhe Institute of Technology (KIT) on efficient data management and analysis, secure data infrastructures, scalable algorithms, and simulation technologies with a view to realizing exascale, machine learning, and data analysis based on artificial intelligence. Another focus is on new computing technologies such as quantum and neuromorphic computing.
Challenges
The amount of data produced in research, technology, industry, and society is constantly growing. This places high demands on the management, security, and analysis of data.
In addition to an excellent infrastructure, such as the necessary computing capacity and sophisticated software, there is a need for experts who can analyse and interpret the data resources from many different disciplines.
Solutions
High-performance computers enable the growing volumes of data to be evaluated in a targeted manner and models to be defined and simulated in detail. All research fields stand to benefit, such as climate research, materials science, astrophysics and particle physics, fluid dynamics, and solid-state physics.
Jülich experts use the excellent infrastructure on hand and the algorithms they have developed to answer questions on topics ranging from civil security and transport to neuroscience and quantum computing. Artificial intelligence (AI) plays an important role here, with Jülich researchers developing and utilizing self-learning programs to gain an even deeper understanding of complex data.
Jülich’s Institute for Advanced Simulation (IAS) is available to provide extensive expertise in all of these areas. The researchers work on a wide variety of topics, developing suitable computational methods, algorithms, and models.
To improve the exchange of information between high-performance computing experts and subject specialists, the Jülich Supercomputing Centre (JSC), which is part of IAS, has established Simulation and Data Laboratories in which experts work together closely on an interdisciplinary basis. They support scientists in working with big data in general as well as in evaluating the data – using methods such as machine learning – and play an active role in their research communities, similar to JSC’s Research Groups.
JSC also conducts interdisciplinary research and development activities in its Algorithms, Tools and Methods Labs, which develop and provide access to generic algorithms, digital tools, and methods in areas such as machine and deep learning, data analysis, mathematical methods and algorithms, and performance analysis. A further important aspect of interdisciplinary activities is support for user communities in the field of high-performance computing.
Infrastructure expansion – for example, between the computing centres in Stuttgart, Garching, and Jülich – not only makes it easier for users to transfer large data sets, but also facilitates the use of multiple centres by greatly reducing data transfer times. This allows users to run simulations on a variety of suitable computer architectures. The three centres work closely together to offer even more advanced data and workflow services. They also develop new features for distributed, collaborative, and real-time visualizations.
Furthermore, IAS operates cross-centre Joint Labs – specialized research units that concentrate on the development of methods and applications for high-performance computers. These Joint Labs create the foundations for the next generation of Earth system models, breakthroughs in brain and materials research, and new, pioneering methods of visualizing scientific data.
To do so, Jülich experts use innovative, data-efficient learning algorithms and new self-supervised learning methods. A major focus of this work is on image processing, computer vision, and machine learning in scientific applications. The application-driven projects range from the segmentation and tracking of cell colonies in microscopic images to the calculation of solar-induced chlorophyll fluorescence from hyperspectral satellite images using artificial neural networks.
Basic research also plays an important role here, with Jülich researchers combining high-performance computing and the most up-to-date theories to gain an understanding of strong interaction phenomena. The phenomena investigated at Jülich range from the dynamics of quarks, gluons, and core systems and strongly correlated electrons in low-dimensional systems right up to collective behaviour in neural networks. Jülich’s most important research activities in this field address questions relating to the origin of life-giving molecules such as carbon and oxygen. Interdisciplinary teams also explore the nature of elements that exist under extreme conditions, such as those found in neutron stars.
Further activities at Jülich focus on the multitude of possible applications for quantum computers, quantum sensors, and quantum communication. The range of such applications is enormous, from secure communication to the planning and operation of digital controls in the energy grid that can reliably be protected from cyber-attacks. To achieve this goal, Forschungszentrum Jülich coordinates the federal state of North-Rhine Westphalia’s project QuGrids, which facilitates value creation at the interface between quantum technologies and energy systems engineering.
Head of the division HPC for Quantum Systems
Head of the Jülich UNified Infrastructure for Quantum computing (JUNIQ)
Group Leader of the Research Group Quantum Information Processing
Professor Quantum Information Processing at RWTH Aachen University
Spokesperson of Helmholtz Information Program 1, Topic 1, and PI in Topics 1 and 2
Deputy director of JSC, head of the division Communication Systems and Services
PI in Helmholtz Information Program 1, Topics 1, 2 and 3 (Topic Board Member)
Head of Jülich Supercomputing Centre
Prof. at Goethe University Frankfurt
Speaker of Helmholtz Information Program 1, PI in Topics 1 and 2 & Joint Lab SMHB
Head of the division HPC for Quantum Systems
Head of the Jülich UNified Infrastructure for Quantum computing (JUNIQ)
Group Leader of the Research Group Quantum Information Processing
Professor Quantum Information Processing at RWTH Aachen University
Spokesperson of Helmholtz Information Program 1, Topic 1, and PI in Topics 1 and 2
Head of RG AI and ML for Healthcare
Full Professor, High Performance Computing & Artificial Intelligence, University of Iceland, Reykjavik, Iceland
PI in Helmholtz Information Program 1, Topic 1 & Joint Lab SMHB
Head of research group Earth System Data Exploration and co-lead of division Large Scale Data Science, University professor in Computational Earth System Science at the University of Cologne
PI in Helmholtz Information Program 1, Topic 1 (Topic Board Member)
