SAMPO：面向生成式世界模型的带运动提示的尺度自回归
SAMPO:Scale-wise Autoregression with Motion PrOmpt for generative world models
AI总结：
本文提出SAMPO生成式世界模型框架，通过尺度自回归、非对称多尺度分词器和轨迹感知运动提示，提升动作条件视频预测与基于模型控制的质量、效率及泛化能力。
摘要：
世界模型使智能体能够在想象环境中模拟动作的后果，以用于规划、控制和长时域决策。然而，现有自回归世界模型因空间结构被破坏、解码效率低下以及运动建模不足，难以产生视觉连贯的预测。为此，我们提出带运动提示的尺度自回归（Scale-wise Autoregression with Motion PrOmpt，SAMPO），这是一种混合框架，将用于帧内生成的视觉自回归建模与用于下一帧生成的因果建模相结合。具体而言，SAMPO将时间因果解码与双向空间注意力相集成，从而保留空间局部性并支持每个尺度内的并行解码。该设计显著提升了时间一致性和展开效率。为进一步增强对动态场景的理解，我们设计了一种非对称多尺度分词器，在观测帧中保留空间细节，并为未来帧提取紧凑的动态表示，从而同时优化内存使用和模型性能。此外，我们引入了轨迹感知运动提示模块，该模块注入关于物体和机器人轨迹的时空线索，将注意力集中于动态区域，并提升时间一致性和物理真实感。大量实验表明，SAMPO在动作条件视频预测和基于模型的控制中取得了具有竞争力的性能，在提升生成质量的同时实现了4.4倍更快的推理。我们还评估了SAMPO的零样本泛化能力和扩展行为，证明了其能够泛化到未见过的任务，并能从更大的模型规模中获益。
英文摘要：
World models allow agents to simulate the consequences of actions in imagined environments for planning, control, and long-horizon decision-making. However, existing autoregressive world models struggle with visually coherent predictions due to disrupted spatial structure, inefficient decoding, and inadequate motion modeling. In response, we propose \textbf{S}cale-wise \textbf{A}utoregression with \textbf{M}otion \textbf{P}r\textbf{O}mpt (\textbf{SAMPO}), a hybrid framework that combines visual autoregressive modeling for intra-frame generation with causal modeling for next-frame generation. Specifically, SAMPO integrates temporal causal decoding with bidirectional spatial attention, which preserves spatial locality and supports parallel decoding within each scale. This design significantly enhances both temporal consistency and rollout efficiency. To further improve dynamic scene understanding, we devise an asymmetric multi-scale tokenizer that preserves spatial details in observed frames and extracts compact dynamic representations for future frames, optimizing both memory usage and model performance. Additionally, we introduce a trajectory-aware motion prompt module that injects spatiotemporal cues about object and robot trajectories, focusing attention on dynamic regions and improving temporal consistency and physical realism. Extensive experiments show that SAMPO achieves competitive performance in action-conditioned video prediction and model-based control, improving generation quality with 4.4$\times$ faster inference. We also evaluate SAMPO's zero-shot generalization and scaling behavior, demonstrating its ability to generalize to unseen tasks and benefit from larger model sizes.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
