RESEARCH PAPERYear 2025

DriveDreamer-2: LLM-Enhanced World Models for Diverse Driving Video Generation

Guosheng Zhao; Xiaofeng Wang; Zheng Zhu; Xinze Chen; Guan Huang; Xiaoyi Bao; Xingang Wang

Classification

Major category: Benchmarks & simulators
Quadrant: Not applicable
Architecture: Not applicable
Prediction paradigm: Not applicable
Subcategories: Neural world simulators
Source review status: Verified from primary sources

Category review. DriveDreamer-2 constructs driving trajectories/layouts and renders multi-camera video with a specialized diffusion model. Its measured downstream application is perception training from synthetic data, not executed driving control. It belongs with learned driving simulators rather than general-purpose components. Reading evidence

AT A GLANCE

Contribution

A contribution summary has not been added yet.

Abstract

An abstract has not been added yet.

Affiliations

Not listed in the collection.

BibTeX

@inproceedings{drivedreamer_2,
  author       = {Guosheng Zhao and
                  Xiaofeng Wang and
                  Zheng Zhu and
                  Xinze Chen and
                  Guan Huang and
                  Xiaoyi Bao and
                  Xingang Wang},
  title        = {DriveDreamer-2: LLM-Enhanced World Models for Diverse Driving Video
                  Generation},
  booktitle    = {AAAI},
  pages        = {10412--10420},
  year         = {2025},

}

Classification

Contribution

Abstract

Affiliations

BibTeX

More in this category

Programmable World Model

FolDeX: A Physical-World Benchmark for Long-Horizon Robotic Manipulation of Deformable Objects

TourPhysics: Bringing Physics to World Models for Exploration and Manipulation from a Single Image