⌁ Awesome World Model Hub Papers · Datasets · Projects
中文 GitHub
← Back to papers

WorldCache: Accelerating World Models for Free via Heterogeneous Token Caching

arXiv 26.3 2026 54.5 method

TLDR

WorldCache accelerates diffusion world models via heterogeneous token caching, achieving 3.7x speedup with 98% rollout quality.

Reasoning

The paper introduces a novel caching framework that addresses token heterogeneity and non-uniform temporal dynamics in diffusion world models, with strong empirical results. However, the approach is tailored to diffusion-based models and may not generalize to other world model architectures.

Read-first score

Read-first score 54.5, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 48.

Recency 6%
100

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

Reproducibility 18%
81

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

Topical relevance 29%
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 18%
50

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

Citation impact 18%
30.2

Uses OpenAlex-shaped citation metadata as a bibliometric attention signal, separate from paper quality. citation_normalized_percentile=0.30229081

Citation velocity 12%
0

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

Field roles

FrontierReproducibility anchor

Rank sensitivity

Stability: volatile; rank range: 539.

Keyword Scores

world model
10
generative world model
8
interactive world model
8
world simulator
7
video world model
6
world dynamics prediction
6
model-based reinforcement learning world model
3

Deep Analysis

Innovations

  • Curvature-guided Heterogeneous Token Prediction using a physics-grounded curvature score and Hermite-guided damped predictor for chaotic tokens
  • Chaotic-prioritized Adaptive Skipping that accumulates a curvature-normalized drift signal and recomputes only when bottleneck tokens drift

Methodology

WorldCache is a caching framework tailored to diffusion world models. It introduces Curvature-guided Heterogeneous Token Prediction, which uses a curvature score to estimate token predictability and applies a Hermite-guided damped predictor for chaotic tokens with abrupt direction changes. It also designs Chaotic-prioritized Adaptive Skipping, which accumulates a curvature-normalized, dimensionless drift signal and recomputes only when bottleneck tokens begin to drift.

Key Results

WorldCache delivers up to 3.7× end-to-end speedups while maintaining 98% rollout quality on diffusion world models.

Resources

PaperProject

Tags