Walk through Paintings: Egocentric World Models from Internet Priors

arXiv 26.1 2026 41.8 method, application

TLDR

EgoWM transforms pretrained video diffusion models into action-conditioned world models using internet priors, enabling accurate future prediction and generalization.

Reasoning

The paper introduces a simple, architecture-agnostic method that repurposes internet-scale video priors for action-conditioned world modeling, demonstrating strong generalization and a new structural consistency metric. However, the abstract lacks explicit real-world benchmarks, and the 'paintings' setting may limit perceived applicability.

Read-first score

Read-first score 41.8, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 56.

Recency 6%

100

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

Topical relevance 29%

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=none

Reproducibility 18%

Screens links and visible text for paper, code, dataset, artifact, and repository signals. pdf=True; code=False; dataset=False; markers=none

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

Frontier

Rank sensitivity

Stability: volatile; rank range: 418.

Keyword Scores

world model

generative world model

video world model

interactive world model

world dynamics prediction

world simulator

model-based reinforcement learning world model

Deep Analysis

Innovations

Transforming any pretrained video diffusion model into an action-conditioned world model via lightweight conditioning layers, enabling controllable future prediction without training from scratch.
Introduction of the Structural Consistency Score (SCS) to evaluate physical correctness independently of visual appearance.
Scaling across diverse embodiments and action spaces, from 3-DoF mobile robots to 25-DoF humanoids, including egocentric joint-angle-driven dynamics.
Demonstration of robust generalization to unseen environments, including navigation inside paintings.

Methodology

EgoWM is a simple, architecture-agnostic method that repurposes the rich world priors of Internet-scale video diffusion models and injects motor commands through lightweight conditioning layers. It requires only modest fine-tuning and scales across embodiments and action spaces, enabling action-conditioned future prediction for navigation and manipulation tasks.

Key Results

EgoWM improves the Structural Consistency Score (SCS) by up to 80% over prior state-of-the-art navigation world models, achieves up to six times lower inference latency, and demonstrates robust generalization to unseen environments, including navigation inside paintings.

Resources

Paper