TUM participates in Mathematical Research Data Initiative
National Research Data Infrastructure for Mathematics
In Germany, universities and research performing organizations (RPO) manage their data differently. The National Research Data Infrastructure (NFDI) aims to network them.
Making research data easier to access and share more effectively. With this goal in mind, the federal and state governments of Germany are funding the development of a National Research Data Infrastructure (NFDI) with 19 consortia for various research areas to date.
In October 2021, the funding for the Mathematical Research Data Initiative (MaRDI) will start. With Professor Mathias Drton from the Department of Mathematics, the Technical University of Munich (TUM) is also involved. He is a co-spokesperson and heads the Statistics and Machine Learning research area together with Professor Bernd Bischl from Ludwig-Maximilians-Universität Munich.
FAIR principle: Research data for all
A lot of research data is only available locally, on a project-by-project basis, and often only for a limited time. The NFDI aims to make these valuable data resources systematically accessible, networked and sustainably usable for the entire German science system.
For transparent and efficient research, the FAIR principles apply: Research data should be (easily) findable, accessible, interoperable, re-usable. To this end, the consortia are developing strategies tailored to different scientific disciplines.
Mathematics consortium MaRDI
Mathematical research data are highly complex, extensive and diverse - whether they are databases of functions, objects, models or algorithms. Moreover, they are widely used in science because mathematics plays an important role in interdisciplinary research.
The Mathematical Research Data Initiative aims to introduce standards for all mathematical research data that will enable certification. To this end, MaRDI is first developing standardized data formats, application programming interfaces (APIs), and query languages. Ultimately, the initiative aims to expand prototypical services and create an infrastructure that will enable all scientific disciplines to safe, access, and use mathematical research data.
Statistics and Machine Learning Task Area
Researchers in Statistics and Machine Learning develop methods to analyze data. These methods enable predictions, facilitate decisions, or reveal structures underlying a scientific phenomenon.
A chief goal of the Statistics and Machine Learning research area is to ensure that conclusions drawn in scientific studies are likely to persist in independent replication studies. Mathematical theory and experimental simulations play together in the processes of method development. Research data range from literature, statistical models and algorithms to benchmark datasets and software.
The research area plans to establish libraries of curated datasets that will be connected to software, research literature, and associated statistical analyses. It will also create workflows and a demonstration platform to evaluate, compare, and align methods through empirical analysis and simulation studies. Finally, the Statistics and Machine Learning Division intends to collaborate with journal partners and establish standards for quality control and reproducibility of numerical experiments in the scientific publication process.
Participants in the Mathematical Research Data Initiative
The MaRDI consortium will be funded for a period of 5 years starting in October 2021. The decision is based on the positive evaluation of the project proposal by a group of experts in a comprehensive review process of the German Research Foundation (DFG).
The Berlin-based Weierstraß Institute for Applied Analysis and Stochastics (WIAS) coordinates MaRDI. The initiative currently includes another 23 partners such as universities and research institutes, scientific infrastructure facilities, clusters of excellence as well as the German Mathematical Society (Deutsche Mathematiker-Vereinigung, DMV) and the European Mathematical Society (EMS). Details can be found on the