AIDE: A federal project for the development of artificial intelligence in Belgium

Federated Learning for Robust, Resilient and Adaptive Protection of Systems

The AIDE project aims to define and implement a federated machine learning platform and demonstrate its effectiveness through several case studies.

Partners

The AIDE project takes place in a rich ecosystem of 4 teams made up of more than 80 researchers

working in the field of AI and cybersecurity.

UCLOUVAIN

Leads the strategic cybersecurity initiative for Wallonia, which brings together all centres and universities, as well as all private players.

IMEC

IDLab from IMEC – Ghent University has been working for many years on decentralizing access to the Internet.

KU Leuven

KU LEUVEN is historically the most active university in AI/Cybersecurity in Belgium.

CETIC

CETIC develops innovative data-sharing technologies and represents Belgium in the large-scale SPARTA project.

Keep up to date with project news

The research team

Summary of human resources involved (each for the entire duration of the project) :

6 PhDs
2 cybersecurity experts
3 international postdoctoral researchers
1 team leader

Miel Verkerken

Ph.D. Researcher (UGent)

Sander Borny

Researcher (IMEC)

Rosana Veroneze

Researcher (UCLouvain)

Khanh Huu The Dam

Ph.D. Researcher (UCLouvain)

Nico Salamone

Research Engineer (CETIC)

Laurens D’hooge

PhD Researcher (Ugent)

Jasper Vaneessen

Researcher (IMEC)

Philippe Massonet

Scientific Coordinator (CETIC)

Davy Preuveneers

Research Manager at DistriNet (KULeuven)

Tim Wauters

Project Manager (IMEC)

Merlijn Sebrechts

Researcher (IMEC)

Wouter Joosen

Head of DistriNet (KULeuven)

Lieven Desmet

Research Manager at DistriNet (KULeuven)

Bruno Volckaert

Professor at IDLab-UGent-imec

Jean Vanderdonckt

Human Computer Interaction (UCLouvain)

Svetla Nikova

Research Manager at COSIC (KULeuven)

Bart Preneel

ProfessorHead of the research group COSIC (KULeuven)

Articles already published or accepted.

Consult our agenda.

Work Packages

We implement our project in two phases. The first phase covers the years 2022 and 2023, according to the dates set out in the call for projects. In the event of a follow-up (i.e., if a second call for projects allows for continuity), a second amplification phase is planned and possible for 2023-2025. The first phase aims to lay the foundations for the work and create the architecture. The second phase aims to amplify and generalize the results and accessibility of the architecture. Our work packages reflect our deliverables according to the two-phase sequence. We have disclosed both sequences to demonstrate the project’s long-term potential and evolving importance.

Five Work Packages (WP) comprise our project following its launch.

WP n°1

Define a federated machine-learning method that respects fundamental rights while ensuring software systems’ resilient and adaptive protection.

WP n°2

Define federated learning-based countermeasures for AI-based attacks.

WP n°3

Define and develop an implementation of a federated learning architecture for AI-based services.

WP n°4

Application of a federated learning architecture for AI-based services.

WP n°5

Transfer and dissemination of research results in the real world.