Human-Centered Health AI

The human-centered health AI group aims to develop and evaluate AI approaches for biomedical discovery and health-related applications. 

Our group aims to advance the state of the art in AI representations of biomolecules such as small molecule metabolites and RNA. Methodologically, we focus on artificial intelligence approaches that combine prior knowledge, often formally represented as ontologies, with large language models, multi-modal models and molecular language models, in order to make knowledge-informed predictions. We also develop and evaluate approaches for enhanced interpretability of such models, and approaches for trustworthiness and safety. 

Our main application areas of expertise are metabolic health, mental health, and evidence synthesis. One of our core areas of interest is the development of approaches that aim to advance knowledge-driven partial automation of biomedical evidence synthesis and thereby to accelerate the translation of evidence into health benefits. We also aim to advance hybrid approaches to integrate multi-modal models with mechanistic predictions, such as for the interpretable and personalised modelling and prediction of personalised metabolism or behavioural health.