Privacy Aware and Privacy Preserving Distributed and Robust Machine Learning for Medical Applications

Project Information

PAROMA-MED

Grant agreement ID: 101070222

DOI

10.3030/101070222

Project closed

EC signature date 22 June 2022

Start date 1 July 2022

End date 30 June 2025

Funded under

• Civil Security for Society

Total cost

€ 4 280 497,50

EU contribution

€ 4 280 497,50

Coordinated by

EURESCOM-EUROPEAN INSTITUTE FOR RESEARCH AND STRATEGIC STUDIES IN TELECOMMUNICATIONS GMBH
Germany

Periodic Reporting for period 2 - PAROMA-MED (Privacy Aware and Privacy Preserving Distributed and Robust Machine Learning for Medical Applications)

Reporting period: 2024-01-01 to 2025-06-30

Summary of the context and overall objectives of the project

The overall objective of the action is to develop, validate and evaluate a platform-based (edge-central cloud) delivery framework for privacy and security assured services and applications in federative cross-border environments. The project work towards achieving this overall objective has been performed in the following way as it is documented in the indicated deliverables.

To this purpose, the project will develop new architectures, technologies, tools and services to support:

automatic attestation of federation partners;
privacy- and security – by-design, integrating standard compliance and performance / QoS requirements into a policy framework;
consumers with their rights for opt-in / opt-out consent, portability and right to be forgotten requests, as well as transparency in access to their private-data;
federative Identity and Access Management, based on Zero Trust principles, continuous risk assessment and on confidentiality, integrity and authenticity insurance;
privacy-preserving and trusted data – storage and – processing in federative environments;
flexible and secure access over the Internet to private-data and service resources;
AI / ML by-design, integrating platform services to be used by application developers for data-intensive applications;
Zero Touch deployment and automatic life-cycle management of services and applications;
managed Privacy and Security operations for automated policy enforcement and cyber-threat detection and mitigation

Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far

Initially, a representative use case was developed with the relevant stakeholders to set the project's overarching goal for integrated demonstration, considering the healthcare specifics. This use case encompasses scenarios involving cross-border interactions and introduces data ownership schemes addressing privacy and data value protection. PAROMA-MED outlined extensive functional and non-functional requirements, recognizing federated ecosystems as integral to the data economy. Furthermore, early-stage analysis of federated machine learning was conducted to support the project's overarching ambition.
Following the definition of requirements and use cases the project embarked on defining functional components and enablers necessary to support use case scenarios. Concepts were proposed and a high-level architecture of the platform has been defined, taking into account the need for early evaluation and validation. Privacy and security frameworks were identified against which project outcomes would be cross-checked to ensure adequate evidence of protection.
PAROMA-MED conducted a deeper analysis of security and privacy frameworks, defining measurable indices for calculating project KPIs.
Concurrently, PAROMA-MED finalised a set of enabling technologies to support cross-border federative practices. Identity Management, based on Zero Trust principles aligned with federated identity management.
A holistic approach was taken to address the data layer of the project platform, focusing on database interactions, data lifecycle management, and traceability and protection mechanisms. The networking level of the project platform was strengthened with the production of the first version of the interconnection prototype, emphasizing strong network mechanisms to ensure maximum protection.
Lastly, use case details were gradually analyzed and tested within a common testbed environment resembling a federation topology, with integration of relevant artifacts underway.

Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)

PAROMA-MED goes beyond the state of the art by delivering a fully integrated, open-source, cloud–edge federative platform that enables secure, privacy-preserving, and compliant cross-border data sharing and AI-driven analytics in healthcare environments. Unlike existing solutions, the platform combines Data Space concepts, zero-trust identity and access management, continuous attestation, consent-aware data governance, federated learning, and DevSecPrivOps into a single, validated framework, demonstrated through real healthcare use cases. Key results include a functional federation marketplace with automated partner attestation and contract enforcement, advanced consent and identity management aligned with GDPR and FHIR standards, and scalable AI services for distributed learning and inference. The project also generated reusable architectural designs, open-source components, and novel IP, contributing to standardisation and interoperability efforts within emerging European Data Spaces. To ensure further uptake and long-term success, key needs include large-scale pilots with healthcare providers, integration into operational health data infrastructures, continued support for standardisation and regulatory alignment, and access to sustainable funding and market mechanisms. These steps will facilitate replication, exploitation, and potential commercialisation of PAROMA-MED results across healthcare and other data-sensitive sectors.