Imagine training a smart medical AI without ever actually seeing or moving the patient records from their original hospital. It's like a teacher visiting different classrooms to learn from students' work without taking the notebooks home. This keeps sensitive health data locked safely in its home country while still allowing the AI to learn how to spot aggressive cancer.
Healthcare providers cannot easily share sensitive patient data across borders due to GDPR and privacy risks. This prevents AI from learning from large, diverse datasets, leading to less accurate cancer diagnoses and costly unnecessary biopsies.
A privacy-enforcing platform for federated AI development. It includes a toolset for diagnosing aggressive prostate cancer and contributions to the HL7 FHIR global data standard.
If you are a software developer dealing with strict GDPR rules and fragmented data — this project developed a federated AI toolset that allows you to train models across 3 different countries without moving sensitive data. This enables the creation of more accurate tools for diagnosing aggressive prostate cancer.
If you are a security firm dealing with the need for secure multi-party computation — this project developed a platform integrating homomorphic encryption and differential privacy. This provides a provenly secure environment for testing and deploying healthcare AI solutions.
If you are a network manager dealing with the high cost of unnecessary biopsies — this project developed a multi-national clinical validation tool. This helps improve predictions of aggressive cancer, reducing patient welfare risks and associated costs.
Based on available project data, no specific pricing or commercial cost models are mentioned; the project is funded by a EUR 6,304,750 EU contribution.
Yes, the project aims to scale up data-driven healthcare internationally by integrating health data hubs from 3 different countries and contributing to the global HL7 FHIR standard.
Based on available project data, specific IP and licensing terms are not detailed, though the consortium includes 4 SMEs and a standards organization.
The project creates guidelines for GDPR-compliant cross-border Federated Learning and includes a legal/ethics partner to ensure regulatory compliance.
The project period runs from 2023-05-01 to 2026-04-30, with the first version of components currently being integrated.
The consortium is well-balanced for commercialization, featuring a 42% industry ratio with 5 industrial partners, including 4 SMEs. With 12 partners across 7 countries, the group combines deep academic research (6 partners) with practical clinical application and a standards organization, ensuring the output meets both technical and regulatory requirements.
Contact the Institut National de Recherche en Informatique et Automatique (INRIA) in France.
Contact us to explore licensing opportunities for the FLUTE privacy-preserving AI toolset.
