标题： PhysVideoGenerator：通过潜在物理引导实现物理感知的视频生成 显示英文标题

标题： PhysVideoGenerator: Towards Physically Aware Video Generation via Latent Physics Guidance

摘要： 当前的视频生成模型可以生成高质量的美学视频，但通常难以学习现实世界物理动态的表示，导致出现不自然的物体碰撞、不一致的重力和时间闪烁等伪影。 在本工作中，我们提出了PhysVideoGenerator，这是一个概念验证框架，它在视频生成过程中显式嵌入可学习的物理先验。 我们引入了一个轻量级预测网络PredictorP，该网络直接从噪声扩散潜在空间中回归从预训练的视频联合嵌入预测架构（V-JEPA 2）中提取的高层物理特征。 这些预测的物理标记通过专用的交叉注意力机制注入到基于DiT的生成器（Latte）的时间注意力层中。 我们的主要贡献是证明了这种联合训练范式的可行性：我们表明扩散潜在空间包含足够的信息来恢复V-JEPA 2的物理表示，并且多任务优化在训练过程中保持稳定。 本报告记录了架构设计、技术挑战和训练稳定性验证，为未来物理感知生成模型的大规模评估奠定了基础。 显示英文摘要

摘要： Current video generation models produce high-quality aesthetic videos but often struggle to learn representations of real-world physics dynamics, resulting in artifacts such as unnatural object collisions, inconsistent gravity, and temporal flickering. In this work, we propose PhysVideoGenerator, a proof-of-concept framework that explicitly embeds a learnable physics prior into the video generation process. We introduce a lightweight predictor network, PredictorP, which regresses high-level physical features extracted from a pre-trained Video Joint Embedding Predictive Architecture (V-JEPA 2) directly from noisy diffusion latents. These predicted physics tokens are injected into the temporal attention layers of a DiT-based generator (Latte) via a dedicated cross-attention mechanism. Our primary contribution is demonstrating the technical feasibility of this joint training paradigm: we show that diffusion latents contain sufficient information to recover V-JEPA 2 physical representations, and that multi-task optimization remains stable over training. This report documents the architectural design, technical challenges, and validation of training stability, establishing a foundation for future large-scale evaluation of physics-aware generative models.