MetaWorld: Scaling Multi-Agent Video World Model from Single-view Video Data
Quick Take
MetaWorld introduces a framework for scaling multi-agent video world models from single-view videos, overcoming challenges of data scarcity and world state alignment. Utilizing Monocular World-State Unrolling and Subject-Aware World Generator, it achieves superior cross-view consistency and identity fidelity, establishing a scalable paradigm for multi-agent video modeling.
Key Points
- Introduces Monocular World-State Unrolling for ego-motion and spatial trajectory decomposition.
- Develops Subject-Aware World Generator for appearance-driven simulation per agent.
- Implements World-State Alignment for consistent physical reality across views.
- Achieves superior cross-view consistency and identity fidelity in experiments.
- Establishes a scalable, physics-driven paradigm for multi-agent video modeling.
Article Content
From source RSS / original summaryarXiv:2606. 02753v1 Announce Type: new Abstract: Video world models are a foundational generative technology for embodied AI and the Metaverse, yet existing approaches are inherently limited to a single agent observing from a single perspective.
Extending these models to multi-agent settings introduces two critical challenges: data scarcity (coordinated multi-view recordings are prohibitively expensive to collect for general open-domain scenarios) and world state alignment (independently generated video streams cannot ensure that shared physical environments and events evolve consistently across views).
To address these challenges, we propose MetaWorld, a novel framework that scales multi-agent video world models to open-domain environments directly from single-view videos. First, we introduce Monocular World-State Unrolling (MWSU) to explicitly decompose monocular footage into the camera operator's ego-motion and the visible subject's spatial trajectory.
This camera-trajectory decomposition naturally extracts synchronized multi-agent motion data within a shared 3D space, completely bypassing the need for multi-camera setups. Second, for precise visual control, we develop the Subject-Aware World Generator to enable appearance-driven simulation conditioned on per-agent identity images.
Finally, to ensure both views are grounded in the identical physical reality, we propose World-State Alignment, a per-frame inter-branch cross-attention mechanism inserted at every transformer layer of the video DiT. By jointly synchronizing the denoising process, WSA enforces both static geometric consistency and dynamic motion consistency, encouraging that the shared 3D environment and physical events remain well-aligned across both egocentric views.
Extensive experiments demonstrate that MetaWorld achieves superior cross-view consistency and identity fidelity, establishing a highly scalable, physics-driven paradigm for multi-agent video world modeling.
Reader Mode unavailable (could not extract clean content).
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.CV
See more →Plan2Map: A Multimodal Benchmark for Document-Grounded Geospatial Boundary Reconstruction from Planning Records
Plan2Map introduces a 208-case benchmark for reconstructing geospatial boundaries from UK planning documents. The GeoPlanAgent system achieves a mean IoU of 0.736, significantly outperforming baseline models, highlighting the challenges in localization and map registration.
