EVEWorld：具身世界模型的物理演化监督
EVEWorld: Physical Evolution Supervision for Embodied World Models
浏览论文内容
中文总结 AI 辅助
EVEWorld通过实例引导恢复和时间实例对齐提供物理演化监督，显著降低模型惰性率，在多个基准上验证了有效性。
中文摘要 AI 辅助
具身世界模型能够为机器人学习实现具身交互的可扩展仿真。然而，现有模型容易产生“模型惰性”，因为它们侧重于视觉保真度而牺牲了物理推理，并且缺乏对操作对象时间动态的过程级监督。在这项工作中，我们提出了EVEWorld，一种用于物理一致目标演化的物理演化监督框架。EVEWorld由两个组件组成：实例引导恢复（IGR）和时间实例对齐（TIA）。首先，IGR通过恢复监督促进实例一致性。其次，TIA通过对齐相邻帧中的目标实例来促进跨帧一致性。我们进一步引入了模型惰性率（MLR），一种衡量生成轨迹中实例一致性持续违反情况的指标。在DreamGenBench、EWMBench和PBench上的大量实验证明了EVEWorld的有效性， notably 与GigaWorld-0相比，MLR降低了87.5%。在WorldArena 2.0 Track 1排行榜上，我们的模型在JEPA相似性方面排名第6，总体排名第17，这进一步验证了我们的演化监督策略的性能。
英文摘要
Embodied world models enable scalable simulation of embodied interactions for robot learning. However, existing models are prone to Model Laziness, as they focus on visual fidelity at the expense of physical reasoning and lack process-level supervision over the temporal dynamics of manipulated objects. In this work, we propose EVEWorld, a physical evolution-supervision framework for physically consistent target evolution. EVEWorld consists of two components: Instance-Guided Restoration (IGR) and Temporal Instance Alignment (TIA). First, IGR promotes instance consistency through restoration supervision. Second, TIA promotes cross-frame consistency by aligning target instances across adjacent frames. We further introduce the Model Laziness Rate (MLR), a metric that measures persistent violations of instance consistency in generated trajectories. Extensive experiments on DreamGenBench, EWMBench, and PBench demonstrate the effectiveness of EVEWorld, notably achieving an 87.5% reduction in MLR compared with GigaWorld-0. On the WorldArena 2.0 Track 1 leaderboard, our model ranks 6th in JEPA Similarity and 17th overall, which further validates the performance of our evolution supervision strategy.
发表机构
- China Merchants Group（招商局集团）
- Nanyang Technological University（南洋理工大学）
机构由 AI 辅助整理，请以论文原文为准。
