PRIVAGAMS will create a platform enabling research institutions to produce high-quality simulated data customized for specific needs, thereby increasing data availability while ensuring privacy remains intact. It will utilize Generative Adversarial Networks (GANs) to create new, simulated yet realistic data samples that replicate the statistical properties of original datasets without incorporating any real data. This approach will encompass various data types, including both tabular, clinical and imaging data. A key challenge is to preserve not only the statistical properties but also the intricate relationships within the data items, such as the link between a medical diagnosis and its corresponding imaging data. To address this, we will provide conditioning capabilities to our generative methods to ensure that these facts are accurately represented in the new data sets.

Cíle udržitelného rozvoje

Masarykova univerzita se hlásí k cílům udržitelného rozvoje OSN, jejichž záměrem je do roku 2030 zlepšit podmínky a kvalitu života na naší planetě.