WorldDreamer: Towards General World Models for Video Generation via Predicting Masked Tokens

TL;DR

WorldDreamer introduces a versatile, large-scale world model for video generation that predicts masked visual tokens, enabling natural scene and driving environment synthesis, and supports multi-modal prompts for interactive tasks.

Contribution

It presents a novel unsupervised visual sequence modeling approach inspired by language models, extending world modeling to general environments beyond specific scenarios.

Findings

Excels in generating diverse videos including natural and driving scenes

Supports tasks like text-to-video, image-to-video, and video editing

Demonstrates versatility and effectiveness in capturing dynamic world elements

Abstract

World models play a crucial role in understanding and predicting the dynamics of the world, which is essential for video generation. However, existing world models are confined to specific scenarios such as gaming or driving, limiting their ability to capture the complexity of general world dynamic environments. Therefore, we introduce WorldDreamer, a pioneering world model to foster a comprehensive comprehension of general world physics and motions, which significantly enhances the capabilities of video generation. Drawing inspiration from the success of large language models, WorldDreamer frames world modeling as an unsupervised visual sequence modeling challenge. This is achieved by mapping visual inputs to discrete tokens and predicting the masked ones. During this process, we incorporate multi-modal prompts to facilitate interaction within the world model. Our experiments show that WorldDreamer excels in generating videos across different scenarios, including natural scenes and driving environments. WorldDreamer showcases versatility in executing tasks such as text-to-video conversion, image-tovideo synthesis, and video editing. These results underscore WorldDreamer's effectiveness in capturing dynamic elements within diverse general world environments.