# EpiWorld: Grounding LLM Policy Agents in Epidemiological World Models

arXiv: https://arxiv.org/abs/2610.02744 · HTML: https://mlsanity.com/paper/2610.02744/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in World models for language agents, one of 11 directions in World models.
- Builds on 17 of its 47 references that are in the map.
- First posted on arXiv 2 Oct 2026.

## Placed in

- World models / [World models for language agents](https://mlsanity.com/wm/world-models-for-language-agents/)

## Builds on (most cited first, 17 of 17)

- [Generative Agents: Interactive Simulacra of Human Behavior](https://mlsanity.com/paper/2304.03442/) (2023)
- [Dream to Control: Learning Behaviors by Latent Imagination](https://mlsanity.com/paper/1912.01603/) (2019)
- [Learning Latent Dynamics for Planning from Pixels](https://mlsanity.com/paper/1811.04551/) (2018)
- [Mastering Diverse Domains through World Models](https://mlsanity.com/paper/2301.04104/) (2023)
- [Mastering Atari with Discrete World Models](https://mlsanity.com/paper/2010.02193/) (2020)
- [Transformers are Sample-Efficient World Models](https://mlsanity.com/paper/2209.00588/) (2022)
- [AgentGym-RL: Training LLM Agents for Long-Horizon Decision Making through Multi-Turn Reinforcement Learning](https://mlsanity.com/paper/2509.08755/) (2025)
- [Multiscale Dynamic Human Mobility Flow Dataset in the U.S. during the COVID-19 Epidemic](https://mlsanity.com/paper/2008.12238/) (2020)
- [Can A Society of Generative Agents Simulate Human Behavior and Inform Public Health Policy? A Case Study on Vaccine Hesitancy](https://mlsanity.com/paper/2503.09639/) (2025)
- [When and How to Lift the Lockdown? Global COVID-19 Scenario Analysis and Policy Assessment using Compartmental Gaussian Processes](https://mlsanity.com/paper/2005.08837/) (2020)
- [Reinforced Epidemic Control: Saving Both Lives and Economy](https://mlsanity.com/paper/2008.01257/) (2020)
- [Where to locate COVID-19 mass vaccination facilities?](https://mlsanity.com/paper/2102.07309/) (2021)
- [Planning Multiple Epidemic Interventions with Reinforcement Learning](https://mlsanity.com/paper/2301.12802/) (2023)
- [AI Agents as Policymakers in Simulated Epidemics](https://mlsanity.com/paper/2601.04245/) (2026)
- [Coordinated Pandemic Control with Large Language Model Agents as Policymaking Assistants](https://mlsanity.com/paper/2601.09264/) (2026)
- [Optimizing Interventions for Agent-Based Infectious Disease Simulations](https://mlsanity.com/paper/2604.02016/) (2026)
- [EpiNarrate: Agentic Generation of Grounded Narratives from Epidemiological Scenario Projections](https://mlsanity.com/paper/2607.15544/) (2026)

## Abstract (arXiv)

Epidemic intervention policies are textual artefacts that human decision-makers interpret, justify, and revise through natural language, making large language models a natural candidate for epidemic policy reasoning. A naive LLM, however, lacks the epidemic dynamics needed to project intervention consequences, the quantitative surveillance signals required to assess severity, and the institutional constraints that define admissible actions. We present EpiWorld, a closed-loop framework that grounds an LLM policy actor in a learned action-conditioned epidemiological world model and a tiered skill library of public-health protocols, surveillance tools, and adaptive lessons accumulated through after-action analysis. Given a candidate intervention, the world model predicts regional epidemic evolution and enables fast counterfactual rollouts that provide feedback for policy selection and refinement. Outcomes of simulated futures are distilled into reusable lessons while protocol constraints remain fixed, allowing the decision process to improve without sacrificing interpretability or controllability. We evaluate both the world model and the end-to-end framework on retrospective COVID-19 and Influenza datasets: the world model achieves the best out-of-distribution Peak-MAE among all forecasting baselines, and the closed-loop framework reduces cumulative hospitalisation by up to 59% across datasets and by an average of ~16% across six LLM backbones, outperforming reinforcement-learning and optimal-control policy baselines.