CausalMotion: Structured Physical Reasoning as Keyframe and Trajectory Guidance for Training-Free Video Generation (opens in new tab)

Recent advances in diffusion-based video generation have significantly improved visual quality and short-term temporal coherence. However, existing methods still struggle to produce videos with physically consistent and causally plausible dynamics, especially in scenarios involving long-horizon interactions. This limitation arises from the fact that video diffusion models primarily learn physical consistency implicitly, while vision-language mod...