发表机构
Stevens Institute of Technology; Northeastern University; Technische Universität Darmstadt（史蒂文斯理工学院; 东北大学; 达姆施塔特工业大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对扩散世界模型去噪推理开销大的问题，提出无需训练的谱缓存框架SpectralCache，利用奇异子空间稳定性与奇异值外推，跳过部分骨干计算，在保持生成质量下实现5.22倍加速。
AI 中文摘要
基于扩散的世界模型能够生成高质量的交互式环境，但由于去噪过程中重复的Transformer评估，导致推理开销巨大。现有的缓存方法主要利用特征或令牌级别的时间冗余，而忽略了扩散特征底层数学结构的探索。在本工作中，我们揭示了世界模型特征在相邻去噪步骤中表现出高度稳定的奇异子空间，而其奇异值遵循可预测的演化模式。基于这一观察，我们提出了SpectralCache，一种无需训练的谱缓存框架，该框架重用稳定的奇异子空间，并通过线性外推仅估计低维奇异值。我们进一步利用相邻全计算特征之间的谱一致性，通过奇异值缩放跳过选定的昂贵骨干网络评估。在代表性世界模型上的大量实验表明，SpectralCache持续提高推理效率，同时保持生成质量。在HunyuanWorld-Voyager-13B上，SpectralCache实现了5.22倍的加速，同时静态场景的WorldScore达到65.90，在推理效率上显著优于现有的无需训练的缓存方法。
英文摘要
Diffusion-based world models enable high-quality interactive environment generation but suffer from substantial inference overhead due to repeated Transformer evaluations during denoising. Existing caching methods mainly exploit temporal redundancy at the feature or token level, leaving the underlying mathematical structure of diffusion features largely unexplored. In this work, we reveal that world-model features exhibit highly stable singular subspaces across nearby denoising steps, while their singular values follow predictable evolution patterns. Building on this observation, we propose SpectralCache, a training-free spectral caching framework that reuses stable singular subspaces and estimates only low-dimensional singular values through linear extrapolation. We further exploit the spectral consistency between neighboring full-computation features to skip selected expensive backbone evaluations via singular value scaling. Extensive experiments on representative world models demonstrate that SpectralCache consistently improves inference efficiency while preserving generation quality. On HunyuanWorld-Voyager-13B, SpectralCache achieves 5.22x acceleration while maintaining a WorldScore of 65.90 for static scenes, substantially outperforming existing training-free caching methods in inference efficiency.
