WorldRover: A Scalable Synthetic Video Data Engine for World Exploration with Rich Annotations

arXiv 2026 41 benchmark, system

TLDR

WorldRover is an Unreal Engine-based synthetic data engine generating richly annotated, long-range video explorations with camera trajectories, geometry, and action signals for world-model learning.

Reasoning

The paper introduces a scalable synthetic data pipeline with dense, aligned annotations and multi-viewpoint rendering, which is a strong contribution. However, it is purely synthetic and the abstract does not report real-world validation or downstream model experiments.

Read-first score

Read-first score 41, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 34.

Recency 6%

100

Uses a gentle age decay so recent papers surface without erasing older foundations. 2026

Methodology quality 18%

Screens visible abstract and analysis fields for experiment, dataset, baseline, metric, and limitation evidence. markers=dataset,experiment,metric,result

Topical relevance 29%

48.6

Uses existing LLM keyword relevance scores normalized to 0-100. world model,world simulator,generative world model,interactive world model,video world model,world dynamics prediction,model-based reinforcement learning world model

Reproducibility 18%

Screens links and visible text for paper, code, dataset, artifact, and repository signals. pdf=True; code=False; dataset=False; markers=dataset

Citation impact 18%

Uses OpenAlex-shaped citation metadata as a bibliometric attention signal, separate from paper quality. cited_by_count=0

Citation velocity 12%

Citation velocity estimates citations per publication-year to reduce old-paper bias. velocity=0.00

Field roles

FrontierMethodology anchor

Rank sensitivity

Stability: volatile; rank range: 413.

Keyword Scores

world model

world simulator

video world model

interactive world model

generative world model

world dynamics prediction

model-based reinforcement learning world model

Deep Analysis

Innovations

A scalable synthetic video data engine (WorldRover-Engine) that generates minute-scale, long-range explorations of artist-built environments while preserving full trajectories and scene geometry.
The engine can replay the same exploration from first-person, third-person, and 360-degree panoramic cameras under different environmental states, including a neutral white material, enabling controlled viewpoint and appearance changes.
The WorldRover-10M dataset pairs RGB with metric depth, camera trajectories, and trajectory-derived action signals; third-person subsets additionally provide dense optical flow, long-range 2D/3D point tracks with visibility, and a separate character trajectory.

Methodology

WorldRover-Engine is an Unreal Engine pipeline that offline-renders minute-scale routes, capturing full camera trajectories and scene geometry. The same exploration can be replayed from multiple camera viewpoints and environmental states. WorldRover-10M is constructed from these renderings, providing richly annotated video sequences.

Key Results

No experimental results are reported in the abstract; the paper focuses on the design of the data engine and the construction of the WorldRover-10M dataset.

Resources

Paper