StableWorld: Towards Stable and Consistent Long Interactive Video Generation

2026-01-21

AI Summary

This work addresses the challenge of error accumulation in interactive video generation during prolonged interactions, which often leads to spatial drift and scene collapse. The study presents the first systematic analysis of the root causes underlying this issue and introduces a model-agnostic dynamic frame culling mechanism that mitigates error propagation at its source by continuously evaluating and filtering out geometrically inconsistent frames in real time. The proposed method integrates seamlessly into prevailing interactive generation frameworks—such as Matrix-Game, Open-Oasis, and Hunyuan-GameCraft—and demonstrates consistent improvements across multiple models in terms of long-term video stability, temporal coherence, and cross-scenario generalization, thereby validating its broad applicability and effectiveness.

Technology Category

• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Computer Vision: Video Understanding & Activity Analysis
• Humans and AI: Game Design — Procedural Content Generation & Storytelling

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

In this paper, we explore the overlooked challenge of stability and temporal consistency in interactive video generation, which synthesizes dynamic and controllable video worlds through interactive behaviors such as camera movements and text prompts. Despite remarkable progress in world modeling, current methods still suffer from severe instability and temporal degradation, often leading to spatial drift and scene collapse during long-horizon interactions. To better understand this issue, we initially investigate the underlying causes of instability and identify that the major source of error accumulation originates from the same scene, where generated frames gradually deviate from the initial clean state and propagate errors to subsequent frames. Building upon this observation, we propose a simple yet effective method, \textbf{StableWorld}, a Dynamic Frame Eviction Mechanism. By continuously filtering out degraded frames while retaining geometrically consistent ones, StableWorld effectively prevents cumulative drift at its source, leading to more stable and temporal consistency of interactive generation. Promising results on multiple interactive video models, \eg, Matrix-Game, Open-Oasis, and Hunyuan-GameCraft, demonstrate that StableWorld is model-agnostic and can be applied to different interactive video generation frameworks to substantially improve stability, temporal consistency, and generalization across diverse interactive scenarios.

Problem

• interactive video generation
• temporal consistency
• stability
• scene collapse
• spatial drift

Innovation

• StableWorld
• temporal consistency
• interactive video generation
• dynamic frame eviction
• error accumulation