Research
Disease development and progression are shaped by complex interactions between genetic, molecular, cellular, physiological, behavioural and environmental factors. As a result, people with the same diagnosis can differ substantially in how their disease progresses and how they respond to treatment. At the same time, biomedical research increasingly relies on diverse data, such as omics, medical imaging, clinical records, laboratory measurements, lifestyle information and patient-reported outcomes, which remain difficult to integrate and interpret together.
A New Approach to Biomedical AI
AIRIS (Mechanism-Informed Multimodal Generative AI for Causal and Dynamical Modelling in Biomedical Research) is developing a new generation of Generative AI specifically designed for biomedical research. By combining complex biomedical data with biological and mechanistic knowledge, AIRIS aims to go beyond conventional, correlation-based AI towards a better understanding of the processes driving disease and uncover novel biomarkers and potential therapeutic targets.
What is Generative AI?
Generative AI refers to AI systems that can create new content or information based on patterns learned from existing data. In biomedical research, this can include generating hypotheses, models, synthetic data or predictions. AIRIS goes a step further and aims to develop a generative AI platform that builds and reasons with mechanistic models of disease.
From Correlations to Mechanisms
A defining feature of AIRIS is its mechanism-informed approach. By integrating Generative AI with causal graphs, dynamic models and biological knowledge, AIRIS will model disease processes across scales – from molecules and cells to tissues, organs and patients. These capabilities come together in the AIRIS platform, which will support biomedical discovery and predictive and personalised medicine. The AI collaborator will support researchers throughout the entire scientific discovery process, enabling them to understand the development of complex diseases and their underlying pathways, and to explore how disease trajectories could change under different interventions.