LSTD: Long Short-Term Temporal Diffusion for Video Generation

arXiv 2026 30.8 method

TLDR

Proposes LSTD, a diffusion model with short-term temporal convolution and long-term temporal attention for temporally consistent video generation and long video inference.

Reasoning

The paper addresses temporal consistency and memory efficiency in video generation, with novel modules and a long-video inference method. Strengths include empirical validation on multiple benchmarks and zero-shot inference; weaknesses are limited to video generation, with no connection to world models or RL concepts.

Read-first score

Read-first score 30.8, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 0.

Recency 5%

100

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

Citation impact 15%

97.2

Uses OpenAlex-shaped citation metadata as a bibliometric attention signal, separate from paper quality. citation_normalized_percentile=0.97190544

Methodology quality 15%

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

Citation velocity 10%

28.5

Citation velocity estimates citations per publication-year to reduce old-paper bias. velocity=3.00

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%

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

Reproducibility 15%

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

Field roles

FoundationFrontier

Rank sensitivity

Stability: volatile; rank range: 591.

Keyword Scores

world model

world simulator

generative world model

interactive world model

video world model

world dynamics prediction

model-based reinforcement learning world model