Technical Partner — Knowledge Graphs & Synthetic Data
We build the Bio-Streams knowledge graph and the synthetic data generation behind the biobank. The graph semantically maps and links heterogeneous health data, and realistic synthetic datasets improve model training, fill gaps in missing data and reduce bias. We also integrate the open toolkit components into the Bio-Streams platform.
Overweight and obesity have reached epidemic levels across Europe, and none of the 53 WHO European Member States is on track to halt the rise by 2025. In children and adolescents the causes are complex, going beyond diet and inactivity. Despite plentiful guidelines, prevention rarely reaches families, schools and clinics effectively.
Bio-Streams creates the first EU Childhood/Adolescence Obesity Biobank, holding real and synthetic data. It feeds a platform offering a knowledge hub, personalised risk assessment and a recommendation engine that draws on the everyday micro-moments that shape behaviour. It is deployed in 7 hospitals and 5 school sites across Europe. Our knowledge graph and synthetic data make the biobank's heterogeneous data interoperable and usable for machine learning.
