MAD：面向高效驾驶世界模型的运动-外观解耦
MAD: Motion Appearance Decoupling for efficient Driving World Models
把视频扩散模型拆成「先预测骨架化运动、再渲染 RGB 外观」两阶段，用极少量监督和算力把通用视频模型改造成可控驾驶世界模型；做视频生成或驾驶仿真的研究者可直接复用这种运动-外观解耦的适配范式与文本/自车/物体控制接口。
原文依据 decouple motion learning from appearance synthesis
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论近期的视频扩散模型能够生成照片级真实、时序连贯的视频，但作为自动驾驶的可靠世界模型仍显不足，因为自动驾驶需要结构化的运动与物理一致的交互。将这些通用视频模型适配到驾驶领域已显示出潜力，但通常需要大量领域特定数据和昂贵的微调。我们提出一种高效的适配框架，以极少的监督将通用视频扩散模型转化为可控的驾驶世界模型。其核心思想是将运动学习与外观合成解耦。首先，模型被适配为以简化形式预测结构化运动：由骨架化智能体与场景元素构成的视频，从而把学习聚焦于物理与社会合理性。随后，复用同一骨干网络，以这些运动序列为条件合成逼真的 RGB 视频，相当于用纹理与光照为运动「穿衣」。这一两阶段过程对应一种推理-渲染范式：先推断动力学，再渲染外观。实验表明，这种解耦方法异常高效：在适配 SVD 时，我们以不足先前 SOTA 模型 6% 的算力达到与之相当的效果。扩展到 LTX 后，我们的 MAD-LTX 模型优于所有开源竞争者，并支持一整套文本、自车与物体控制。项目主页：https://vita-epfl.github.io/MAD-World-Model/
英文原文摘要 1477 字符
Recent video diffusion models generate photorealistic, temporally coherent videos, yet they fall short as reliable world models for autonomous driving, where structured motion and physically consistent interactions are essential. Adapting these generalist video models to driving domains has shown promise but typically requires massive domain-specific data and costly fine-tuning. We propose an efficient adaptation framework that converts generalist video diffusion models into controllable driving world models with minimal supervision. The key idea is to decouple motion learning from appearance synthesis. First, the model is adapted to predict structured motion in a simplified form: videos of skeletonized agents and scene elements, focusing learning on physical and social plausibility. Then, the same backbone is reused to synthesize realistic RGB videos conditioned on these motion sequences, effectively "dressing" the motion with texture and lighting. This two-stage process mirrors a reasoning-rendering paradigm: first infer dynamics, then render appearance. Our experiments show this decoupled approach is exceptionally efficient: adapting SVD, we match prior SOTA models with less than 6% of their compute. Scaling to LTX, our MAD-LTX model outperforms all open-source competitors, and supports a comprehensive suite of text, ego, and object controls. Project page: \href https://vita-epfl.github.io/MAD-World-Model/ https://vita-epfl.github.io/MAD-World-Model/