World Models for Policy Refinement in StarCraft II

TL;DR

This paper introduces StarWM, a world model for StarCraft II that predicts future observations, enabling a decision system that significantly improves AI performance in complex, partially observable environments.

Contribution

We develop StarWM, the first world model for SC2 with structured textual representations and a new dataset, enhancing policy refinement through a generate-simulate-refine loop.

Findings

StarWM achieves nearly 60% better resource prediction accuracy.

StarWM improves self-side macro-situation consistency.

StarWM-Agent increases win rates by up to 30% against advanced AI levels.

Abstract

Large Language Models (LLMs) have recently shown strong reasoning and generalization capabilities, motivating their use as decision-making policies in complex environments. StarCraft II (SC2), with its massive state-action space and partial observability, is a challenging testbed. However, existing LLM-based SC2 agents primarily focus on improving the policy itself and overlook integrating a learnable, action-conditioned transition model into the decision loop. To bridge this gap, we propose StarWM, the first world model for SC2 that predicts future observations under partial observability. To facilitate learning SC2's hybrid dynamics, we introduce a structured textual representation that factorizes observations into five semantic modules, and construct SC2-Dynamics-50k, the first instruction-tuning dataset for SC2 dynamics prediction. We further develop a multi-dimensional offline evaluation framework for predicted structured observations. Offline results show StarWM's substantial gains over zero-shot baselines, including nearly 60% improvements in resource prediction accuracy and self-side macro-situation consistency. Finally, we propose StarWM-Agent, a world-model-augmented decision system that integrates StarWM into a Generate--Simulate--Refine decision loop for foresight-driven policy refinement. Online evaluation against SC2's built-in AI demonstrates consistent improvements, yielding win-rate gains of 30%, 15%, and 30% against Hard (LV5), Harder (LV6), and VeryHard (LV7), respectively, alongside improved macro-management stability and tactical risk assessment.