Pub-AI: AI in Science

Driving · 2026-09-30

Graph World Models for Constrained Epidemic Policy Planning

Yiqi Su, Rashed Shelim, Lingyi Wang, Walid Saad, Naren Ramakrishnan

arXiv:2609.35545PDF

Auto-summarized: this summary was generated by a language model from the paper’s text and has not been reviewed by an editor. Check the paper before relying on it. Benchmark numbers appear only after a human has verified them.

TL;DR

EpiMind is a graph world-model framework for multi-region epidemic policy planning under shared-resource constraints. It uses a graph-factored recurrent state-space model (GF-RSSM) to generate joint policy-conditioned rollouts, and a graph-temporal ADMM (GT-ADMM) planner to coordinate region interventions, enforce per-period feasibility via projection, and evaluate temporal specifications on projected actions.

Why it matters

The authors target epidemic planning where regions are coupled by mobility and share limited resources, and where planners need action-conditioned prediction plus per-period feasibility guarantees. EpiMind combines graph-structured latent dynamics with constrained receding-horizon planning so candidate actions can be rolled out under a learned model, projected onto a feasible shared-resource set, and verified with temporal-logic robustness after projection.

Method

  • Graph-factored recurrent state-space model (GF-RSSM): parameter-shared transition with region-specific beliefs/states, using graph attention to exchange mobility-weighted neighbor information; trained offline and frozen during planning.
  • Joint policy-conditioned rollouts: future observations unavailable, so imagined trajectories use the learned prior recursively; dedicated heads decode health quantities used by the planner.
  • Graph-temporal constrained planning (GT-ADMM): alternates region action optimization, shared-resource projection onto a feasible set (encoded linear box and sum constraints), coordination updates, then rerolls projected actions and evaluates temporal specifications.

Limitation

The paper notes that under matched intervention burden, graph coordination provides only a measurable but incremental benefit.

Abstract (from arXiv)

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.

Related papers