In an increasing digitalisation era, the availability of data is growing exponentially in many varied areas: science, business, media, etc. Conventional data management procedures and evaluation techniques become obsolete and cannot cope any longer with the amount of data and its speed acquisition. At the same time, new data science approaches offer unprecedented opportunities to search for patterns and extract information from data. Within the DLR, these new data management techniques are already being used in different areas. Therefore, the mission of this department is to extend this work and to provide support in the field of new data management and data analysis techniques. Currently the department is divided into four main areas: Data Management Technologies, Climate Informatics, Visual Analytics and Machine Learning.
Data Management Technologies
The ubiquitous flood of generated data places enormous demands on efficient data access, data management and data archiving. In this context, the main research areas of the working group include the development of new methods and processing strategies for large, heterogeneous, multidimensional data sets (e.g. from Earth observation and radio astronomy) in distributed IT infrastructures, such as public/private computing infrastructures. ...[more]
A better understanding of the climate system is essential not only for the progress of basic science, but also to assess the risks of unabated climate change. The Climate Informatics group develops modern data science methods based on causal inference, deep learning, and nonlinear time series analysis, and closely collaborates with climate scientists to address this great challenge. ...[more]
The Visual Analysis working group is concerned with the research and development of novel visual methods for the analysis of large and complex data, which place humans and their cognitive abilities more at the centre of the analysis process. ...[more]
The Machine Learning (MLE) group is focused on the development of novel methods of machine learning, in particular deep learning. These methods are detached from the concrete applications, but are oriented towards the needs of DLR. ...[more]