Medical world models: representing medical states, modelling clinical dynamics and guiding intervention policies
A new review paper proposes the development of "medical world models" to advance healthcare AI beyond static diagnoses. These models aim to create internal simulators of patient-state dynamics, enabling clinicians to anticipate disease progression and compare intervention outcomes. The paper outlines a roadmap focusing on patient-state construction, clinical dynamics modeling, and intervention decision support to achieve these goals. AI
IMPACT This research could lead to AI systems that provide more dynamic and predictive insights into patient health trajectories.
- artificial intelligence
- arXiv
- reinforcement learning
- foundation model
- digital twin
- Treatment Effect Estimation Using Nonlinear Two‐Stage Instrumental Variable Estimators: Another Cautionary Note
- Medical world models
- Longitudinal modelling of body mass index from birth to 14 years.
- Disease simulation in medical images
- patient-state construction
- clinical dynamics modelling
- intervention decision support
- perception--dynamics--planning systems