assignments

The sub-project of Jena University Hospital in the AVATAR project (AVATAR TV-UKJ) opens up medical data pools for AVATAR that are already publicly accessible and additionally need to be integrated. It supplements the architecture of the AVATAR data pool with components for federated data use. The subproject develops special anonymization approaches by generating synthetic data using AI models. In addition, AVATAR TV-UKJ designs and implements methods for training AI models on exclusively decentrally managed datasets (Privacy Preserving Deep Learning) and integrates them into the AVATAR infrastructure. In order to promote the willingness to donate data and to use AVATAR methods in data utilization projects, the subproject develops information and training offers tailored to the target group.