环强制法：面向自回归视频扩散模型的精准长期记忆
Ring Forcing: Towards Precise Long-Term Memory for Autoregressive Video Diffusion
浏览论文内容
中文总结 AI 辅助
针对自回归视频扩散模型的长期记忆瓶颈，提出Ring Forcing框架，通过环形训练策略、压缩与时间步组合策略及稀疏RoPE机制，实现分钟级视频生成的优异连贯性与物体恒常性，性能优于现有方法。
中文摘要 AI 辅助
将视频生成长时间序列时会暴露出一个关键瓶颈：当前模型缺乏稳健的长期记忆。该缺陷可从两个关键维度研究：一是物体恒常性，即物体重新出现时精准复现其外观的能力；二是记忆容量，即处理超长上下文并利用远距历史信息的能力。稳健的长期记忆需要同时具备这两种能力：仅靠物体恒常性而上下文处理不足会限制时间范围，仅靠长上下文长度而缺乏恒常性则无法保持物体身份。为解决此问题，本文提出Ring Forcing，一种旨在稳健构建和精准利用长期记忆的自回归视频扩散框架。其环形结构训练策略强制从远距历史中检索信息，有效平衡了严格历史一致性与生成多样性之间的权衡；为扩展记忆容量，引入了压缩与时间步组合策略，在固定序列长度约束下，该方法将有效历史范围延长至分钟级时长，并对整个历史实现全面感受野；此外，还提出稀疏RoPE机制以实现灵活、可扩展的记忆适配，同时充分利用预训练先验。大量实验表明，Ring Forcing实现了优异的分钟级连贯性和物体恒常性，显著优于现有最先进方法。
英文摘要
Scaling video generation to long durations reveals a critical bottleneck: current models lack robust long-term memory. This deficiency can be studied along two critical aspects: object permanence, the ability to precisely reproduce the appearance of objects upon re-entry; and memory capacity, the ability to process ultra-long context and use information from distant history. Robust long-term memory requires both: object permanence without sufficient context handling limits the temporal scope, while long context length without permanence fails to maintain identity. To address this, we present Ring Forcing, an autoregressive video diffusion framework designed to robustly construct and precisely utilize long-term memory. Our ring-structured training strategy enforces retrieval from distant history, effectively reconciling the trade-off between strict historical adherence and generative diversity. To expand memory capacity, we introduce a compression and timestep composition strategy. Under fixed sequence length constraints, this method extends the effective historical span to minutes-long durations and achieves a comprehensive receptive field over the entire history. Furthermore, we present a sparse RoPE mechanism to enable flexible, scalable memory adaptation while fully exploiting pre-trained priors. Extensive experiments demonstrate that Ring Forcing achieves superior minutes-long coherence and object permanence, significantly outperforming state-of-the-art methods.
发表机构
- Stanford University（斯坦福大学）
- Massachusetts Institute of Technology（麻省理工学院）
- Peking University（北京大学）
- University of California, Berkeley（加州大学伯克利分校）
- ByteDance（字节跳动）
机构由 AI 辅助整理，请以论文原文为准。
