Toward Stable World Models: Measuring and Addressing World Instability in Generative Environments
This work addresses the issue of world instability in generative environments, which can introduce noise into agent learning and compromise performance in safety-critical settings, representing an incremental improvement in world modeling.
The paper tackles the problem of preserving previously generated scenes over time in diffusion-based world models, which is crucial for applications like reinforcement learning and game engines, by introducing an evaluation framework that measures world stability and revealing significant challenges in achieving high stability.
We present a novel study on enhancing the capability of preserving the content in world models, focusing on a property we term World Stability. Recent diffusion-based generative models have advanced the synthesis of immersive and realistic environments that are pivotal for applications such as reinforcement learning and interactive game engines. However, while these models excel in quality and diversity, they often neglect the preservation of previously generated scenes over time--a shortfall that can introduce noise into agent learning and compromise performance in safety-critical settings. In this work, we introduce an evaluation framework that measures world stability by having world models perform a sequence of actions followed by their inverses to return to their initial viewpoint, thereby quantifying the consistency between the starting and ending observations. Our comprehensive assessment of state-of-the-art diffusion-based world models reveals significant challenges in achieving high world stability. Moreover, we investigate several improvement strategies to enhance world stability. Our results underscore the importance of world stability in world modeling and provide actionable insights for future research in this domain.