[1035] DS4DH - Data Science for Digital Health

The Data Science for Digital Health (DS4DH) group is part of the Department of Radiology and Medical Informatics at the Faculty of Medicine, University of Geneva. 

Led by Prof. Douglas Teodoro and located at Campus Biotech, the group develops computational methods to manage and analyze complex datasets in the health and life sciences. The research integrates artificial intelligence, natural language processing, machine learning and medical informatics to transform health data into digital health solutions.

The research program is organized into three core streams. The first is biomedical natural language processing, which focuses on the design of methods for semantic text understanding and information extraction to transform unstructured records into computable data. The second stream is deep patient representation, which designs deep learning-based algorithms, such as transformers and graph neural networks, to model longitudinal patient trajectories and dynamic changes. The third stream leverages these foundation approaches for clinical safety and risk assessment, in which we build predictive systems to model disease dynamics and automate risk assessment in clinical research.

To validate our algorithms in real-world clinical settings, the group collaborates closely with the University Hospitals of Geneva (HUG) and Swiss private actors in the health and life sciences sectors.