Hierarchical Denoising For Multi-Step Visual Reasoning

arXiv 2026 21.4 method

TLDR

Proposes HDR, a hierarchical denoising framework for multi-step visual reasoning in video generation, achieving improved reasoning consistency and low latency.

Reasoning

Strengths: novel hierarchical latent structure enabling coarse-to-fine reasoning, a new benchmark with six diverse tasks, and significant performance gains over baselines. Weaknesses: limited to synthetic visual reasoning tasks, no explicit connection to real-world video data or scalability analysis.

Read-first score

Read-first score 21.4, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 0.

Recency 6%

100

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

Methodology quality 18%

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

Reproducibility 18%

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

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

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: 37.

Keyword Scores

world model

world simulator

generative world model

interactive world model

video world model

world dynamics prediction

model-based reinforcement learning world model

Resources

Paper