NS-Diff：流体纳维-斯托克斯引导的强化学习视频扩散
NS-Diff: Fluid Navier-Stokes Guided Video Diffusion via Reinforcement Learning
把简化的 Navier-Stokes 约束与最小加加速度原理写成可微/策略梯度的奖励，用来微调视频扩散模型，为「生成结果要符合物理」这类需求提供了可复用的 RL 后训练范式；其噪声鲁棒的运动区域检测与物理条件潜变量注入也可直接迁移到其他条件视频生成任务。
原文依据 a reinforcement learning optimization module that enforces simplified Navier-Stokes constraints on fluid dynamics and minimum-jerk principles on rigid bodies through policy gradients
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论尽管近期的视频生成模型取得了令人印象深刻的视觉质量，但生成物理上合理的视频仍然具有挑战性，尤其是对流体动力学和刚体运动而言。为解决这一问题，我们提出 NS-Diff，一个面向视频扩散的物理引导强化学习框架。首先，我们设计了一个噪声鲁棒的物理动力学检测器，通过分析含噪潜变量帧中的运动来区分刚体区域与流体区域。其次，我们引入物理条件潜变量注入模块，该模块对速度场、形变梯度和材质掩码进行编码，并通过交叉注意力将其注入 DiT 去噪器。第三，我们引入强化学习优化模块，通过策略梯度对流体动力学施加简化的纳维-斯托克斯约束，对刚体施加最小加加速度原理。在 PhysVideoBench、UCF 和 MSR-VTT 上的实验表明，我们的方法将加加速度误差降低 43%，将流体散度降低 33%，并将 FVD 改善 22.7%，实现了更高的物理合理性与视觉质量。
英文原文摘要 1069 字符
While recent video generation models achieve impressive visual quality, generating physically plausible videos remains challenging, especially for fluid dynamics and rigid-body motions. To address this, we present **NS-Diff**, a physics-guided reinforcement learning framework for video diffusion. First, we design a noise-robust physical dynamics detector that distinguishes rigid and fluid regions by analyzing motion in noisy latent frames. Second, we introduce a Physics-Conditioned Latent Injection module, which encodes velocity fields, deformation gradients, and material masks, and injects them into the DiT denoiser via cross-attention. Third, we introduce a reinforcement learning optimization module that enforces simplified Navier-Stokes constraints on fluid dynamics and minimum-jerk principles on rigid bodies through policy gradients. Experiments on PhysVideoBench, UCF, and MSR-VTT show that our approach reduces jerk errors by 43%, decreases fluid divergence by 33%, and improves FVD by 22.7%, achieving higher physical plausibility and visual quality.