Imagine wanting to find a cure for a disease by looking at patient records from ten different hospitals, but none of the hospitals are allowed to share their data due to privacy laws. Instead of moving the data to one place, this technology sends the AI model to the data, learns from it locally, and only shares the 'lessons learned' without ever seeing the actual records. It adds extra locks and shields to ensure no one can reverse-engineer the private information from those lessons.
Organizations cannot share high-quality datasets for AI research because of GDPR and strict internal access policies. Current Federated Learning methods are still vulnerable to data leakage and re-identification attacks.
The Armored Federated Learning (AFL) platform and a certification tool for GDPR compliance in AI implementations.
If you are a cancer hospital dealing with siloed patient data and strict GDPR rules — this project developed the Armored Federated Learning (AFL) platform that allows AI-driven insights from cross-border datasets while ensuring patient privacy.
If you are a software firm dealing with vulnerabilities like inference attacks or curious aggregators in AI — this project developed privacy enhancement methods including Homomorphic Encryption and Differential Privacy that exceed GDPR requirements.
If you are a compliance firm dealing with the difficulty of certifying AI privacy — this project developed a tool and metric for the certification of GDPR compliance for Federated Learning implementations.
Based on available project data, no specific pricing or cost structures are mentioned.
Yes, the project objective is to deliver a highly scalable Federated AI service platform designed for multi-site, cross-domain, and cross-border datasets.
Based on available project data, specific IP ownership and licensing terms are not provided.
The platform is designed to enable GDPR compliance and provides privacy guarantees that exceed the requirements of GDPR, including a specific tool for certification.
The project period is from 2022-10-01 to 2025-12-31.
The consortium is well-balanced for commercialization, featuring 10 partners across 7 countries. With a 30% industry ratio (including 3 SMEs), the project blends academic research from 1 university and 5 research organizations with practical clinical application from 2 clinical partners, ensuring the technology is tested in real-world medical environments.
Contact FUNDACION CENTRO TECNOLOXICO DE TELECOMUNICACIONS DE GALICIA
Contact us to explore licensing the Armored Federated Learning platform for your healthcare data needs.
