This project explores how artificial intelligence (AI), and in particular machine learning, can be used to model robust and accurate disease biomarkers, measurable indicators that guide diagnosis and treatment decisions. To enable this, we have developed the BIODAI toolbox, a computational framework designed to build and optimize machine learning–based biomarker models with high predictive power and robustness. While many ML-based biomarker models remain untested outside research environments, this project will evaluate the BIODAI toolbox in real-world settings. We will test it in hospitals, to help doctors predict which cancer patients are most likely to benefit from immunotherapy, and in industry, to support the development of health solutions for liver disease. Beyond technical validation, the project will also assess how such ML-based biomarker models can be translated into practice, examining their regulatory readiness and usability in both clinical and industrial contexts.

2026

2027