
This research was conducted in collaboration with Wiam Belouard and Sara El Kardi as part of the 4th-year AI Research Clinics in aivancity’s Grande École Program, under the supervision of Dr. Antoun Yaacoub, a professor at aivancity and Director of the MSc in Generative Artificial Intelligence.
The paper introduces X-SHIELD, an innovative approach to explainability-by-design tailored for self-healing systems orchestrated by multi-agents.
Selected students will join aivancity’s Grande École Program while benefiting from an enhanced curriculum that includes 200 hours of additional training per year.
Today, the most advanced artificial intelligence systems are capable of automatically detecting anomalies, correcting errors, or adjusting their behavior without human intervention. However, these mechanisms often remain opaque to users and operators, turning the decisions made by AI into veritable "black boxes."
In response to this major challenge regarding transparency, students at aivancity have developed a solution capable of instantly generating a structured and understandable explanation of the decisions made by the system.
X-SHIELD offers two levels of interpretation:
This research was officially published in the proceedings of the ENASE 2026 conference (Volume 1, pages 662–669, ISBN: 978-989-758-828-0), thereby permanently establishing the students’ work in the international scientific literature.
The acceptance of the paper at ENASE, a renowned international conference ranked CORE B in software engineering and computer science, serves as significant validation of the project’s methodological and scientific quality.
In addition to being published, the work was also selected for an oral presentation at the conference, once again highlighting the excellence of the work produced.
This project fully embodies aivancity’s educational philosophy: to train professionals who can master cutting-edge technologies while taking into account the ethical, human, and organizational challenges of artificial intelligence.