Graph based model improves epidemic planning across regions with shared resources
Graph World Models for Constrained Epidemic Policy Planning
Machine LearningArtificial Intelligence
Summary
Planning epidemic responses is hard because different regions affect each other and resources like hospital beds are limited. The authors created EpiMind, a computer model that represents regions as connected points on a graph to predict how diseases spread and how policies impact each area. It also helps find plans that follow resource limits by adjusting strategies over time. Their method made better predictions and planned effective responses that respected shared resource limits.
What this means in practice
- •For public health planners: Coordinate interventions across regions using a method that respects shared limited medical resources and predicts disease spread more accurately.
- •For urban mobility analysts: Model how movement between city areas influences epidemic dynamics to optimize region-specific policies within resource limits.
Authors
Yiqi Su, Rashed Shelim, Lingyi Wang, Walid Saad, Naren Ramakrishnan
Abstract
Epidemic policy planning often requires coordination between geographical regions, taking into account mobility-driven spillovers and how to make use of limited resources. Existing methods either lack action-conditioned models of coupled dynamics or cannot guarantee per-period feasibility. We present EpiMind, a graph world model framework for constrained epidemic policy planning across regions. A graph-factored recurrent state-space model generates joint policy-conditioned rollouts from regional latent beliefs, while graph-temporal ADMM optimizes regional interventions, enforces shared-resource feasibility through projection, and evaluates temporal specifications under the learned model. EpiMind reduces admission RMSE by 29% relative to graph-free dynamics modeling, plans within 1-5% of the best feasible constant policy with guaranteed shared-budget feasibility, and outperforms all deployable baselines across three resource budgets in real-context evaluation. These results demonstrate that graph-structured policy imagination with explicit constrained coordination supports effective epidemic interventions from learned dynamics.