发表机构
Georgia Institute of Technology; Lawrence Berkeley National Lab; KTH Royal Institute of Technology; Nordita; International Computer Science Institute（佐治亚理工学院; 劳伦斯伯克利国家实验室; 瑞典皇家理工学院; 北欧理论物理研究所; 国际计算机科学研究所）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
变分流（VSF）通过学习条件潜在分布，在物理时间中实现高效概率预测，兼具流匹配的灵活性与流（SF）的计算效率，在长时程和分叉动力学中表现优越，并可集成到JEPA世界模型提升任务成功率。
AI 中文摘要
概率预测对于预测复杂动力系统非常重要，因为内在随机性和不完整的观测可能导致相同的观测状态演化为多个可能的未来。虽然流匹配是概率预测的一种灵活方法，但计算成本高昂。流（SF）通过直接在物理时间中学习连续速度场，重新表述了这种方法以高效建模时间演化。然而，SF学习的是确定性速度场。因此，对于给定的固定初始状态和观测历史，它只能提供单一的未来轨迹。为了克服这一限制，我们引入了变分流（VSF）。我们的方法学习一个以感兴趣的动力为条件的潜在分布。这反过来使得概率预测成为可能。重要的是，我们通过在物理时间中生成来保持SF的计算效率。在确定性和随机动力系统中，VSF展示了优越的预测准确性和分布保真度。我们展示了在超过1000步的长时程滚动以及具有分叉动力学的设置中的优势。此外，VSF可以作为即插即用的预测器集成到现有的基于联合嵌入预测架构（JEPA）的世界模型中，以提高导航、运动规划和操作中的时间动力学和目标导向成功率。
英文摘要
Probabilistic forecasting is important for predicting complex dynamical systems because intrinsic randomness and incomplete observations can cause the same observed state to evolve into multiple plausible futures. While flow matching is a flexible approach for probabilistic forecasting, it is computationally expensive. Streaming flow (SF) reformulates this approach to model temporal evolution efficiently by learning a continuous velocity field directly in physical time. However, SF learns a deterministic velocity field. Thus, it provides only a single future trajectory for a given fixed initial state and observation history. To overcome this limitation, we introduce Variational Streaming Flow (VSF). Our approach learns a latent distribution that is conditioned on the dynamics of interest. In turn, this enables probabilistic forecasting. Importantly, we retain the computational efficiency of SF by generating in physical time. Across deterministic and stochastic dynamical systems, VSF demonstrates superior predictive accuracy and distributional fidelity. We demonstrate the advantage for both long-horizon rollouts exceeding 1,000 steps, and settings with bifurcating dynamics. Moreover, VSF can be integrated into existing Joint-Embedding Predictive Architecture (JEPA)-based world models as a plug-and-play predictor to improve temporal dynamics and goal-directed success rate in navigation, motion planning, and manipulation.
