← Catalog
PaperWayve · 2025
GAIA-2: A Controllable Multi-View Generative World Model for Autonomous Driving
Latent diffusion world model generating controllable, multi-camera driving video from structured inputs.
Summary
Presents GAIA-2, a latent diffusion world model for driving that generates high-resolution, spatiotemporally consistent multi-camera video conditioned on ego dynamics, agents, road semantics, and environmental factors — enabling scalable simulation of common and rare scenarios across the UK, US, and Germany.
Metadata
- Authors
- Wayve
- Year
- 2025
- arXiv
- 2503.20523
- Introduces
- GAIA-2
Relationships
Introduces
#driving#diffusion#multi-view