Can World and Action Modeling Benefit Each Other under Unified Causal Training?

arXiv 2026 36.5 method

TLDR

OneWM unifies causal training of world prediction and action generation in video-action architectures, showing reciprocal benefits in LIBERO, RoboCasa365, and real-world policy experiments.

Reasoning

The paper offers a strong empirical comparison across simulated and real-world embodied environments, directly testing bidirectional improvements between world modeling and action learning. Its weakness is that the abstract does not discuss architectural limitations or failure cases, so the generality of the reciprocal benefits remains unclear.

Read-first score

Read-first score 36.5, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 48.

Recency 5%

100

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

Topical relevance 25%

68.6

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 15%

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

Reproducibility 15%

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

Graph prestige 15%

5.7

PageRank-style prestige over the local citation graph built from available OpenAlex references. local_pagerank=0.0016

Citation impact 15%

Uses OpenAlex-shaped citation metadata as a bibliometric attention signal, separate from paper quality. cited_by_count=0

Citation velocity 10%

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

Field roles

Frontier

Rank sensitivity

Stability: volatile; rank range: 416.

Keyword Scores

world model

world dynamics prediction

video world model

interactive world model

generative world model

model-based reinforcement learning world model

world simulator

Deep Analysis

Innovations

Introduces OneWM, a unified causal training framework that treats actions as both conditions for world prediction and targets for policy learning within a shared video-action architecture.
Provides a direct empirical investigation showing reciprocal benefits between action-conditioned world modeling and action learning.

Methodology

The paper proposes OneWM, a unified causal training framework using a shared video-action architecture in which actions condition world prediction and are also generated as policy targets. It evaluates policy learning and world modeling on LIBERO, LIBERO-Plus, and RoboCasa365, and complements these with real-world policy experiments.

Key Results

Learning visual consequences of specified actions improves policy performance and generalization across changing scenes, while action learning improves multi-view world-prediction fidelity under the same architecture.

Resources

Paper