FadeMem: 面向自回归视频扩散的距离感知记忆巩固
FadeMem: Distance-Aware Memory Consolidation for Autoregressive Video Diffusion
AI总结：
提出FadeMem，一种距离感知的KV记忆巩固机制，在固定缓存预算下将历史KV块组织成时间层次，通过幂律分配实现近密远疏的记忆，提升长视频生成中的主体一致性和时间连贯性。
摘要：
自回归视频生成器通过生成连续的时间片段来合成长时间视频，但其历史KV缓存随视频长度增长。现有的有界缓存方法通过局部窗口、汇合令牌或压缩记忆状态来降低这一成本，但它们通常为历史的不同部分分配固定角色。我们提出FadeMem，一种距离感知的KV记忆巩固机制，在固定缓存预算下将历史KV块组织成时间层次。这一设计受频率依赖的时间衰减启发：细节快速去相关，而粗略场景结构和身份在更长时间内保持有用。在生成过程中，新历史作为细粒度条目插入，而较旧的相邻条目在幂律时间分配调度下逐步合并，从而在单个缓存中产生近密远疏的记忆。无需架构更改，FadeMem保留近期上下文以处理短期动态，并保留紧凑的长程锚点以保持身份和场景连贯性。实验表明，与现有有界缓存策略相比，主体一致性、背景稳定性和时间连贯性均得到提升。
英文摘要：
Autoregressive video generators synthesize long videos by generating successive temporal segments, but their historical KV cache grows with video length. Existing bounded-cache methods reduce this cost with local windows, sink tokens, or compressed memory states, yet they usually assign fixed roles to different parts of the history. We propose FadeMem, a distance-aware KV memory consolidation mechanism that organizes historical KV blocks into a temporal hierarchy under a fixed cache budget. This design is motivated by frequency-dependent temporal decay: fine details decorrelate quickly, while coarse scene structure and identity remain useful over longer horizons. During generation, new history is inserted as fine-grained entries, while older adjacent entries are progressively merged under a power-law temporal allocation schedule, yielding a dense-near, sparse-far memory within one cache. Without architectural changes, FadeMem improves long-range consistency while largely preserving visual quality, and lightweight adaptation further enhances motion dynamics and visual fidelity. Using the same unified schedule under a fixed cache budget, FadeMem also remains effective over multi-minute and hour long video generation and reduces peak memory under a matched KV budget.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
