Humboldt-Universität zu Berlin - Statistics

As part of the funding initiative in the field of artificial intelligence (AI), the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) is funding the new joint research group DeSBi of the Humboldt-Universität zu Berlin, the Hasso Plattner Institute (HPI), the University of Potsdam (UP), the Charité - Universitätsmedizin Berlin, the Max Delbrück Center for Molecular Medicine and the Fraunhofer Heinrich Hertz Institute: In the future, this research group wants to provide biomedical scientists with better methods for gaining knowledge from structured data. The spokesperson of the research group is Sonja Greven and Nadja Klein is a Principal Investigator of the research group.

 

In the research group, experts from the fields of machine learning and statistics will work together to improve interpretability, uncertainty quantification and statistical inference for deep learning and to improve the modelling flexibility of statistical methods for structured data. In particular, methods will be developed to enable statistical inference for structured data through uncertainty quantification, hypothesis testing and adjustment for confounding variables, and to improve explanations of structured data through hybrid statistical and deep learning models, population and distribution-level explanations, and robust sparse explanations. There will be a feedback loop between method development and biomedical applications, as they collect structured data such as images, gene sequences or history data and require quantification of uncertainty, adjustment for confounding variables and testing of hypotheses with statistical error control, in addition to good predictions. Moreover, the methods developed will be applicable in many fields.