At the ESHG – European Society of Human Genetics, one of the most influential forums in genomics and precision medicine, the conversation is clearly shifting.
It’s no longer just about data availability. It’s about how to use it: securely, collaboratively and at scale.
Within this context, the Better AI Health project contributed with a poster on federated, multimodal analysis for rare diseases, showcasing how AI can operate across institutions without moving sensitive data.
Why this matters:
- Healthcare data is fragmented by design
- Privacy constraints are non-negotiable
- Yet, clinical value depends on combining datasets
Federated learning is where these constraints turn into an opportunity.
Our work shows that it is already possible to:
- integrate clinical, genomic, and imaging data across centers
- train AI models without centralizing data
- improve diagnostic accuracy in complex and rare diseases
