An AI model (ICCV 2023) for zero-shot text-to-video generation using image diffusion models. It allows generating short video clips from text prompts without training on video data.