Forcing-KV: Hybrid KV Cache Compression for Efficient Autoregressive Video Diffusion Models

arXiv 2026 16.8 method

TLDR

Forcing-KV compresses KV caches in autoregressive video diffusion models via hybrid static/dynamic attention head pruning, reducing memory and speeding generation.

Reasoning

The paper provides a clear empirical analysis of attention head specialization and proposes a practical compression method with measurable speedups and memory savings. However, the abstract is incomplete and lacks explicit quality metrics, and the connection to world models is not established.

Read-first score

Read-first score 16.8, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 3.

Recency 5%

100

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

Reproducibility 15%

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

Methodology quality 15%

Screens visible abstract and analysis fields for experiment, dataset, baseline, metric, and limitation evidence. markers=none

Graph prestige 15%

5.7

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

Topical relevance 25%

4.3

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 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: 27.

Keyword Scores

world model

generative world model

video world model

world simulator

interactive world model

world dynamics prediction

model-based reinforcement learning world model

Resources

Paper