My team’s ultimate goal is to create explainable and robust AI models for multimodal multi-centre medical data. Our research is deeply rooted in three key areas:

• explainable AI, which seeks to make the reasoning of AI algorithms transparent and understandable;
• privacy-preserving machine learning, aimed at developing techniques that safeguard patient identity while improving the quality of prediction models;
• and out-of-distribution generalization, which ensures that our AI models remain accurate and reliable even faced with data originating from different populations and acquisition settings.