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

A Survey: Learning Embodied Intelligence from Physical Simulators and World Models

arXiv 25.07 2025 63.7 survey

TLDR

A survey on integrating physical simulators and world models for embodied intelligence, bridging simulation and real-world deployment.

Reasoning

The paper provides a comprehensive review of two key technologies for embodied AI, but as a survey it lacks original experiments or real-world evaluations. Its strength lies in synthesizing recent advances, though it does not introduce new methods or datasets.

Read-first score

Read-first score 63.7, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 42.

Recency 8%
86.7

Uses a gentle age decay so recent papers surface without erasing older foundations. 2025

Methodology quality 25%
80

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

Topical relevance 42%
60

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 25%
46

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

Field roles

FrontierMethodology anchor

Rank sensitivity

Stability: volatile; rank range: 179.

Keyword Scores

world model
10
world dynamics prediction
8
world simulator
7
model-based reinforcement learning world model
7
generative world model
5
interactive world model
4
video world model
1

Deep Analysis

Innovations

  • Systematic review of recent advances in learning embodied AI through integration of physical simulators and world models
  • Analysis of complementary roles of physical simulators and world models in enhancing autonomy, adaptability, and generalization
  • Synthesis of current progress and identification of open challenges for embodied AI systems
  • Discussion of interplay between external simulation and internal modeling to bridge sim-to-real gap

Methodology

This survey systematically reviews recent literature on embodied intelligence, focusing on the integration of physical simulators and world models. It analyzes their complementary roles in robot learning, discusses the interplay between external simulation and internal modeling, and synthesizes progress while identifying open challenges.

Key Results

The survey highlights that physical simulators and world models enhance autonomy, adaptability, and generalization in intelligent robots, and discusses strategies for bridging the gap between simulated training and real-world deployment.

Limitations

  • The survey does not present new experimental results or quantitative comparisons
  • Coverage may be limited by the rapidly evolving nature of the field and potential selection bias in reviewed works
  • Open challenges are identified but not empirically validated

Resources

Paper

Tags