Imagine you need a huge library of patient records to train a medical AI, but the books are locked in different vaults across Europe for privacy. Instead of trying to move the real books, this project creates 'fake' but statistically identical copies that look and act like real data. This allows researchers to study rare blood diseases without ever seeing a real patient's private identity.
Medical research for rare blood diseases is stalled because patient data is fragmented across isolated hospitals and locked by strict privacy laws. This makes it nearly impossible to get the large datasets required to train accurate AI models.
A federated computing infrastructure and AI pipelines that generate synthetic patient data and anonymize real records while keeping data local.
If you are a drug discovery firm dealing with a lack of patient data for rare diseases—this project developed AI pipelines for synthetic data generation that allow you to train models on realistic datasets without violating GDPR.
If you are an AI developer dealing with fragmented data silos in hospitals—this project developed a federated computing infrastructure that lets you train algorithms across multiple sites without moving raw sensitive data.
If you are a clinic dealing with the inability to share rare disease data due to strict privacy laws—this project developed anonymization and synthetic data tools that enable secure cross-border research collaboration.
Based on available project data, no specific pricing or commercial cost is mentioned; however, the project outcomes are intended to be made openly available to healthcare and industry players.
The project establishes a cross-border federated computing infrastructure connecting 10 countries, indicating a design intended for large-scale European deployment.
Based on available project data, the project aims to make pipelines, standards, and data openly available to stakeholders in healthcare, academia, and industry.
It uses federated learning, secure multiparty computation, and differential privacy to ensure data is processed in a privacy-preserving fashion that complies with European regulations.
The project is active from 2022-12-01 to 2026-11-30.
The consortium is well-balanced for technology transfer, consisting of 16 partners across 10 countries. With a 31% industry ratio (5 companies, including 2 SMEs), there is significant commercial interest in the output. The mix of 6 universities and 5 research centers ensures the technical depth of the AI and privacy protocols is matched by practical industrial application.
Contact Universidad Politecnica de Madrid
Contact us to explore licensing the synthetic data pipelines for your rare disease research.
