DataNord

Overview

Besides "in vivo" data from actual use, "in vitro" data is also obtained experimentally in countless laboratory tests. By using models , previously trained via machine learning, synthetic "in silico" data can now also be produced. By generating this type of synthetic data, both new data sets can be produced and gaps in existing data sets can be filled post hoc. As part of the data competence center DataNord, the SynData project explores these new approaches for generating research data in respect to modeling the quality of synthetic datasets, extending the existing models of the structure of datasets and augmenting these black-box methods for generating synthetic datasets with explanatory components.