DreamWorld: Unified World Modeling in Video Generation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Tan, Boming, Zhang, Xiangdong, Liao, Ning, Zhang, Yuqing, Zhang, Shaofeng, Yang, Xue, Fan, Qi, Zhang, Yanyong
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917300648542208
author Tan, Boming
Zhang, Xiangdong
Liao, Ning
Zhang, Yuqing
Zhang, Shaofeng
Yang, Xue
Fan, Qi
Zhang, Yanyong
author_facet Tan, Boming
Zhang, Xiangdong
Liao, Ning
Zhang, Yuqing
Zhang, Shaofeng
Yang, Xue
Fan, Qi
Zhang, Yanyong
contents Despite impressive progress in video generation, existing models remain limited to surface-level plausibility, lacking a coherent and unified understanding of the world. Prior approaches typically incorporate only a single form of world-related knowledge or rely on rigid alignment strategies to introduce additional knowledge. However, aligning the single world knowledge is insufficient to constitute a world model that requires jointly modeling multiple heterogeneous dimensions (e.g., physical commonsense, 3D and temporal consistency). To address this limitation, we introduce \textbf{DreamWorld}, a unified framework that integrates complementary world knowledge into video generators via a \textbf{Joint World Modeling Paradigm}, jointly predicting video pixels and features from foundation models to capture temporal dynamics, spatial geometry, and semantic consistency. However, naively optimizing these heterogeneous objectives can lead to visual instability and temporal flickering. To mitigate this issue, we propose \textit{Consistent Constraint Annealing (CCA)} to progressively regulate world-level constraints during training, and \textit{Multi-Source Inner-Guidance} to enforce learned world priors at inference. Extensive evaluations show that DreamWorld improves world consistency, outperforming Wan2.1 by 2.26 points on VBench. Code will be made publicly available at \href{https://github.com/ABU121111/DreamWorld}{\textcolor{mypink}{\textbf{Github}}}.
format Preprint
id arxiv_https___arxiv_org_abs_2603_00466
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle DreamWorld: Unified World Modeling in Video Generation
Tan, Boming
Zhang, Xiangdong
Liao, Ning
Zhang, Yuqing
Zhang, Shaofeng
Yang, Xue
Fan, Qi
Zhang, Yanyong
Computer Vision and Pattern Recognition
Despite impressive progress in video generation, existing models remain limited to surface-level plausibility, lacking a coherent and unified understanding of the world. Prior approaches typically incorporate only a single form of world-related knowledge or rely on rigid alignment strategies to introduce additional knowledge. However, aligning the single world knowledge is insufficient to constitute a world model that requires jointly modeling multiple heterogeneous dimensions (e.g., physical commonsense, 3D and temporal consistency). To address this limitation, we introduce \textbf{DreamWorld}, a unified framework that integrates complementary world knowledge into video generators via a \textbf{Joint World Modeling Paradigm}, jointly predicting video pixels and features from foundation models to capture temporal dynamics, spatial geometry, and semantic consistency. However, naively optimizing these heterogeneous objectives can lead to visual instability and temporal flickering. To mitigate this issue, we propose \textit{Consistent Constraint Annealing (CCA)} to progressively regulate world-level constraints during training, and \textit{Multi-Source Inner-Guidance} to enforce learned world priors at inference. Extensive evaluations show that DreamWorld improves world consistency, outperforming Wan2.1 by 2.26 points on VBench. Code will be made publicly available at \href{https://github.com/ABU121111/DreamWorld}{\textcolor{mypink}{\textbf{Github}}}.
title DreamWorld: Unified World Modeling in Video Generation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2603.00466