New article in “Information Science” and award for ZMI-work

Award for holistic generative AI workflow for the secure use of medical data

Our research assistant at ZMI, Waldemar Hahn, was honored with the Sax-FDM OPEN DATA Award, the jury's special prize, for his work in the field of generative artificial intelligence. The award was given for a comprehensive generative AI workflow that covers the entire process from data preparation to the selection of suitable generative models, hyperparameter tuning, model training, and the selection of high-quality synthetic data.

The aim of this workflow is to make it much easier to share medical data – in a way that complies with data protection regulations, is quality-assured, and is tailored to the target group. The synthetic data sets generated can be provided in various harmonized formats and are therefore suitable for different user groups (e.g., research, teaching, development). This improves access to clinical data without revealing sensitive original data.

The work is part of Waldemar Hahn's ongoing dissertation project entitled “Model-Agnostic Quality Improvement for Synthetic Tabular Clinical Data Generation and Publication.” In it, he pursues a model-agnostic approach to improving the quality of synthetic, tabular clinical data, which systematically enhances the comparability, reusability, and publishability of such data.

In addition to the award, the relevant scientific article on the topic was accepted for publication in the journal “Information Sciences” and published on December 15. The publication underscores the scientific relevance and practical benefits of the approach:

➡️ https://www.sciencedirect.com/science/article/pii/S0020025525010631

Special thanks for their support go to Dr. Markus Wolfien and Prof. Dr. Martin Sedlmayr.

We congratulate on this award and the successful publication and are delighted about the important contribution to the responsible use of generative AI in medicine.