⌁ Awesome World Model Hub Papers · Datasets · Projects
中文 GitHub
← Back to papers

X-World: Controllable Ego-Centric Multi-Camera World Models for Scalable End-to-End Driving

arXiv 26.3 2026 69.9 method, system

TLDR

X-World is an action-conditioned multi-camera generative world model for scalable end-to-end autonomous driving simulation.

Reasoning

The paper introduces a novel simulator that generates future multi-view video streams conditioned on actions and optional controls, addressing limitations of real-world testing. Strengths include explicit cross-view consistency and controllability; weaknesses include potential lack of real-world validation details in the abstract.

Read-first score

Read-first score 69.9, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 58.

Recency 8%
100

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

Topical relevance 42%
82.9

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

Methodology quality 25%
70

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

Reproducibility 25%
38

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

Field roles

FrontierMethodology anchor

Rank sensitivity

Stability: volatile; rank range: 147.

Keyword Scores

world model
10
world simulator
10
generative world model
10
video world model
9
world dynamics prediction
8
interactive world model
7
model-based reinforcement learning world model
4

Deep Analysis

Innovations

  • Action-conditioned multi-camera generative world model that simulates future observations directly in video space
  • Controllable over dynamic traffic agents and static road elements, with a text-prompt interface for appearance-level control (e.g., weather, time of day)
  • Video style transfer by conditioning on appearance prompts while preserving underlying action and scene dynamics
  • Multi-view latent video generator designed to explicitly encourage cross-view geometric consistency and temporal coherence under diverse control signals

Methodology

X-World is a multi-view latent video generator that takes synchronized multi-view camera history and a future action sequence to generate future multi-camera video streams. It incorporates optional controls over dynamic agents and static elements, and a text-prompt interface for appearance-level control. The model is designed to encourage cross-view geometric consistency and temporal coherence under diverse control signals.

Key Results

X-World achieves high-quality multi-view video generation with strong view consistency across cameras, stable temporal dynamics over long rollouts, and high controllability with strict action following and faithful adherence to optional scene controls.

Resources

Paper

Tags