Prisma-World: Camera-Controllable Multi-Agent Video World Model
2026-06-08 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors present Prisma-World, a video world model designed to create consistent videos from multiple agents viewing the same scene. Unlike earlier models that generate each agent's video independently, Prisma-World jointly generates all agent views together, using camera positions and geometry to ensure that overlapping views match in objects and layout. They also introduce a new dataset, PrismaDataset, which provides diverse multi-agent video data with detailed camera info to help train and test such models. Their experiments show Prisma-World can produce high-quality, controllable multi-agent videos that stay consistent across different viewpoints.
video world modelsmulti-agent generationcross-view consistencycamera conditioningdenoising processrelative camera geometryRoPE (Rotary Position Embeddings)curriculum trainingminimap guidanceUnreal Engine 5 dataset
Authors
Huiqiang Sun, Zhan Peng, Size Wu, Kun Wang, Kang Liao, Dianyi Wang, Xingyu Zeng, Sheng Jin, Yangguang Li, Zhiguo Cao, Ziwei Liu, Wei Li
Abstract
Video world models have made rapid progress in generating controllable visual experiences, but most of them still simulate the world from a single observer. Extending such models to multiple agents raises a central challenge: if each agent's future state is generated independently, overlapping views may instantiate different versions of the same scene, leading to inconsistent objects, layouts, and appearances across agents. Conventional camera conditioning controls individual trajectories, but it does not explicitly couple the generation of views that should agree under shared scene geometry. We introduce Prisma-World, a camera-controllable multi-agent world model that formulates multi-agent generation as a joint geometry-aware denoising process for cross-view consistency. Prisma-World processes all agent videos within one full-attention sequence, uses a multi-agent RoPE design to distinguish agent identities while preserving synchronized temporal coordinates, and injects relative camera geometry into attention to bias overlapping viewpoints toward shared scene evidence. To further strengthen multi-view consistency and enhance global spatial perception, we augment our framework with an overlap-decaying curriculum training paradigm alongside minimap-conditioned structural guidance. To facilitate the training and evaluation of multi-agent models, we introduce PrismaDataset, a large-scale UE5 dataset with panoramic acquisition across diverse scenes, composable multi-agent view groups with flexible agent counts and complex camera trajectories, and precise camera/action annotations for consistency training and evaluation. Experiments show that a single Prisma-World model can generate high-fidelity multi-agent videos with flexible agent numbers, camera controllability, improved cross-view consistency, and spatial grounding under minimap guidance.