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

From Generative Engines to Actionable Simulators: The Imperative of Physical Grounding in World Models

arXiv 26.1 2026 66 survey

TLDR

Argues world models must be physically grounded actionable simulators, not just video generators, using medical decision-making as a test case.

Reasoning

Strengths: Clearly identifies limitations of current world models (visual conflation) and proposes a reframing towards causal structure and constraints. Weaknesses: As a survey, it may lack novel empirical contributions; the medical stress test is mentioned but not detailed in abstract.

Read-first score

Read-first score 66, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 60.

Recency 8%
100

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

Topical relevance 42%
85.7

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%
50

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

Reproducibility 25%
38

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

Field roles

Frontier

Rank sensitivity

Stability: volatile; rank range: 358.

Keyword Scores

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

Deep Analysis

Innovations

  • Reframing world models as actionable simulators rather than visual engines, emphasizing physical grounding and causal structure.
  • Proposing structured 4D interfaces, constraint-aware dynamics, and closed-loop evaluation as key components.
  • Using medical decision-making as an epistemic stress test to demonstrate the necessity of counterfactual reasoning and intervention planning.

Methodology

This survey analyzes current world models, showing they fail under intervention and violate invariant constraints despite high-fidelity video generation. It proposes a reframing toward actionable simulators with structured 4D interfaces, constraint-aware dynamics, and closed-loop evaluation, using medical decision-making as a case study to illustrate the requirements.

Key Results

Modern world models frequently violate invariant constraints, fail under intervention, and break down in safety-critical decision-making, demonstrating that visual realism is an unreliable proxy for world understanding.

Resources

Paper

Tags