World Model for Robot Learning: A Comprehensive Survey

arXiv 2026 63.3 survey

TLDR

A comprehensive survey of world models for robot learning, covering architectures, applications, datasets, and future directions.

Reasoning

The paper provides a thorough and systematic review of world models in robot learning, effectively bridging fragmented literature and highlighting key paradigms and applications. However, as a survey, it does not introduce novel experimental results or methods, which limits its direct contribution beyond synthesis.

Read-first score

Read-first score 63.3, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 57.

Recency 6%

100

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

Topical relevance 29%

81.4

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 18%

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

Citation impact 18%

69.6

Uses OpenAlex-shaped citation metadata as a bibliometric attention signal, separate from paper quality. citation_normalized_percentile=0.69557445

Reproducibility 18%

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

Citation velocity 12%

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

Field roles

FrontierBridgeMethodology anchor

Rank sensitivity

Stability: volatile; rank range: 430.

Keyword Scores

world model

video world model

world dynamics prediction

generative world model

model-based reinforcement learning world model

world simulator

interactive world model

Deep Analysis

Innovations

Comprehensive review of world models from a robot-learning perspective, addressing fragmented literature
Taxonomy of how world models are coupled with robot policies, as learned simulators, and as video generation models
Connection of world model concepts to navigation and autonomous driving domains
Summary of representative datasets, benchmarks, and evaluation protocols

Methodology

The authors conducted a systematic literature review of world models in robot learning, examining architectures, functional roles, and applications. They categorized works based on how world models are coupled with policies, as learned simulators for reinforcement learning and evaluation, and as robotic video world models progressing from imagination-based generation to controllable, structured, and foundation-scale formulations. The review also connects these ideas to navigation and autonomous driving, and summarizes representative datasets, benchmarks, and evaluation protocols.

Key Results

The survey identifies key paradigms in world models for robot learning, including policy coupling, learned simulation, and video generation. It also provides a summary of representative datasets, benchmarks, and evaluation protocols, and highlights major challenges and future directions for predictive modeling in embodied agents.

Resources

Paper