DecMem: 面向分钟级一致世界生成的解耦记忆方法

摘要

近期视频生成模型的进展推动了可控世界模型的快速发展。然而，在长程推理下保持细粒度的时空一致性仍是一个关键挑战。本文突破显式三维记忆与粗粒度帧级隐式建模的局限，提出一种细粒度、可学习且可扩展的记忆机制，用于实现一致的世界生成。我们首先识别了朴素可学习记忆架构在长程外推中的两个根本限制：计算效率低下与注意力分散。通过对注意力分散的系统性分析，我们提出 DecMem——一种解耦式记忆架构，采用稀疏全局记忆实现全局历史的高效细粒度访问，并借助锚定局部记忆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的方法。通过确保精确高效的长时记忆并实现卓越的外推能力，DecMem 能够以高保真度与一致性生成分钟级可控长视频。

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.