Abstract
World model synthesis aims to convert interaction experiences into an internal model of environment dynamics. Existing symbolic approaches often fit observed transitions or mixtures of local rules but do not produce a complete executable program that can operate independently of the real environment. We present Mind-Studio, a framework that synthesizes executable pygame-style world models from state-action-next-state trajectories using large language models.
Mind-Studio combines entropy-selected traces with a lightweight game skill file that contains object, action, and static scene information extracted from screenshots. We evaluate synthesis quality with a K-step lookahead fidelity protocol that compares generated world-model rollouts against Real-ALE rollouts from the same state. On Montezuma's Revenge, Mind-Studio improves chosen-action next-state prediction from 0.3% for PoE-World to 48.7%, verifying 5 of 8 subgoals; across Alien, Assault, and Skiing, it achieves stronger branch-level fidelity than prior learned lookahead sources.
Blogger's Review: Mind-Studio's innovation lies in its ability to combine large language models with game dynamics synthesis, significantly enhancing behavior prediction accuracy in partially observable games. The framework holds immense potential for applications in training and testing agents in complex environments, making it a noteworthy development in AI model creation and game design that could lead to a new wave of technological advancements.