发表机构
University of Science and Technology of China (USTC); Suzhou Institute for Advanced Research, USTC（中国科学技术大学; 中国科学技术大学苏州高等研究院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对现实世界关系型环境的动态演化，提出图世界建模（GWM）框架，并构建基准GWM-Zero，设计WorldGraph模型，结合状态感知图变换器与转换感知GRPO，在三种转换粒度上超越现有基线。
AI 中文摘要
世界模型推断环境的潜在状态，以捕捉其底层动态并预测未来演化。然而，许多现实世界环境本质上是关系型的，并作为演化图被观测，其中实体、关系及其属性随时间变化。先前的图相关世界模型使用图结构来组织内部状态或支持特定任务的推理，而非将演化图本身视为被建模的世界。我们转而研究图世界建模（GWM），其中图演化本身构成世界动态。我们针对观测到的图演化、潜在图状态和异构图转换预测来公式化图世界建模。基于这一公式，我们构建了GWM-Zero，一个覆盖八个时序图数据集上节点级、边级和图级转换的基准。我们提出WorldGraph，它结合了用于多粒度结构和转换条件演化建模的状态感知图变换器，以及使用动态分组和结构感知可验证奖励的转换感知GRPO。在GWM-Zero上的大量实验表明，WorldGraph在所有三种转换粒度上均一致优于代表性的图表示、时序图学习、图预训练和图世界模型基线。
英文摘要
World models infer latent states of an environment to capture its underlying dynamics and predict future evolution. Many real-world environments, however, are inherently relational and observed as evolving graphs, where entities, relations, and their properties change over time. Prior graph-related world models use graph structures to organize internal states or support task-specific reasoning, rather than treating an evolving graph itself as the modeled world. We instead study graph world modeling (GWM), where graph evolution itself constitutes the world dynamics. We formulate graph world modeling over observed graph evolution, latent graph states, and heterogeneous graph-transition predictions. Based on this formulation, we construct GWM-Zero, a benchmark covering node-, edge-, and graph-level transitions over eight temporal graph datasets. We propose WorldGraph, which combines a state-aware graph transformer for multi-granularity structural and transition-conditioned evolution modeling with transition-aware GRPO using dynamic grouping and structure-aware verifiable rewards. Extensive experiments on GWM-Zero show that WorldGraph consistently outperforms representative graph representation, temporal graph learning, graph pretraining, and graph world-model baselines across all three transition granularities.
