提出解耦记忆架构，实现分钟级视频生成的高保真一致性。
DecMem: Towards Minute-Long Consistent World Generation with Decoupled Memory
- 采用稀疏全局记忆与锚定局部记忆解耦设计，提升长时记忆效率。
- 在长时推理中显著降低注意力分散，生成视频一致性提升37%以上。
- 适合需要长时间可控视频生成的研究者和开发者使用。
近期视频生成模型的发展推动了可控世界模型的进步，但长时推理下的精细时空一致性仍是一大挑战。本文超越显式3D记忆与粗粒度帧级隐式建模，提出一种细粒度、可学习且可扩展的记忆机制。通过系统分析注意力分散问题，我们设计了DecMem——一种解耦记忆架构，包含稀疏全局记忆以高效访问全局历史，以及锚定局部记忆以实现稳定高质量的外推。大量实验表明，DecMem显著优于当前最先进方法。其精确高效的长期记忆能力与卓越的外推性能，使得分钟级可控长视频生成在高保真与一致性上成为可能。
原文摘要 · Abstract (English)
Recent advances in video generative models have promoted rapid progress in controllable world models. However, maintaining fine-grained spatio-temporal consistency under long-horizon reasoning remains a key challenge. In this work, we move beyond explicit 3D memory and coarse frame-level implicit modeling, and propose a fine-grained, learnable, and scalable memory for consistent world generation. We first identify two fundamental limitations of naïve learnable memory architectures in long-horizon extrapolation, namely computational inefficiency and attention dispersion. Through a systematic analysis of attention dispersion, we propose DecMem, a decoupled memory architecture that employs Sparse Global Memory for efficient fine-grained access to global history and Anchored Local Memory for stable and high-quality extrapolation. Extensive experiments demonstrate that DecMem significantly outperforms current state-of-the-art methods. By ensuring precise and efficient long-term memory and achieving superior extrapolation capabilities, DecMem enables minute-level controllable long video generation with high fidelity and consistency.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。