SECURED InnoHub Concept

In SECURED, we aim to create and manage a Privacy-enhancing Hub (the SECURED InnoHub) that will provide tools, services, and overall support to external involved third parties of the healthcare domain, including researchers, Innovators or Health Data users as well as EU data Hubs across Europe, thus facilitating them to perform accurate data analytics in a distributed and private matter. The SECURED hub is also meant to promote collaboration among parties by acting as a one stop collaboration point for involved parties to share results and collaboratively shape/enhance their expertise through the Hub with others in a privacy-preserving matter. Given that the dominant health data analytics underlined technology is Machine Learning and Deep Learning intelligence, we focus our data analytics tools and services on enhancing the privacy of ML/DL solutions by offering a secure multiparty computation-capable toolbox that can operate in various modes under a SECURED Federation infrastructure. The end goal of the SECURED hub is to bring together providers and consumers of health data and offer them a trusted, secure and privacy-preserving environment to research, test their solutions and to collaborate.

Synthetic data generation

We will leverage unbiased Federated Learning (FL) AI-based methods, such as generative adversarial networks and representation learning, to extract features and distributions from real data and produce synthetic data with similar features and distribution.

Data anonymization

We will combine statistical and semantic techniques for anonymizing data to minimize unwanted data leakage while ensuring the utility of the data for legitimate purposes (e.g., scientific research).

Secure Multiparty computation

Leveraging recent breakthroughs in cryptography and in bryptograhpic engineering, we will develop new hardware and software based approaches to address the current limitations of multiparty computation in terms of performance and computational overhead.