DuoMatching: Joint-Marginal Distribution Matching for Few-Step Video Generation

arXiv 2026 21.2 method

TLDR

Proposes DuoMatching, a joint-marginal distribution matching framework for few-step video generation, improving visual quality and semantic alignment via frame-level supervision from an image generator.

Reasoning

The paper presents a novel method for video generation with strong experimental validation, including human evaluations. However, its core focus is distribution matching for video synthesis, not explicitly world modeling or dynamics prediction, so relevance to world-model keywords is limited.

Read-first score

Read-first score 21.2, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 6.

Recency 5%

100

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

Methodology quality 15%

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

Reproducibility 15%

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

Topical relevance 25%

8.6

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

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: 127.

Keyword Scores

video world model

world dynamics prediction

world model

generative world model

world simulator

interactive world model

model-based reinforcement learning world model

Resources

Paper