TU 5: Data-intensive modeling
[H. Dreiner, M. Kramer, M. Griebel]
In recent years, technological developments have led to new and often unforeseen challenges in almost all fields of science, from biomedicine to physics and astronomy. Our ability to collect data exceeds our ability to validate but also analyze, visualize, store, and curate all the information. But besides these challenges that are naturally addressed in other TUs, the large data sets also allow for a fundamental change in doing science. In what is sometimes called “the fourth paradigm”, one does not compare data to existing first-principle models anymore, but one can use the data to select which among competing models is favored by the data. CASCADE combines researches working in diverse fields, who have in common that they all face big data challenges in order to make scientific progress. Examples therefore include streaming of large amounts of data and real-time processing using exascale computing (e.g. for the creation of manifolds for radio frequency interference mitigation) or applications in complex networks in biology. The TU “Data Intensive modeling” will establish a hub that allows us to develop and adopt methods in areas that will also go beyond the initial set-up of CASCADE, aiming to cause a fundamental change in scientific exploration.
