提出概率框架，让多个扩散模型生成更协调。
SyncSDE: A Probabilistic Framework for Diffusion Synchronization
- 用概率建模多轨迹相关性，替代盲目拼接分数。
- 不同任务采用最优相关模型，效果优于统一方法。
- 为协同生成提供理论依据，适合多模型协作场景。
已有研究尝试通过融合多个扩散模型的得分来协同生成，扩展至原始领域之外。主流方法通过混合估计得分以人工关联生成过程，但现有方法依赖如平均等简单启发式策略，未考虑任务特异性。这些方法未能解释为何有效，且当适用于某一任务的启发式被盲目应用于其他任务时，常导致次优结果。本文提出一种概率框架，分析扩散同步的原理，揭示启发式应聚焦的环节：建模多轨迹间的相关性，并根据具体任务进行适配。我们进一步识别出各任务的最优相关模型，在多个任务上均优于此前在所有任务上使用单一启发式的做法。
原文摘要 · Abstract (English)
There have been many attempts to leverage multiple diffusion models for collaborative generation, extending beyond the original domain. A prominent approach involves synchronizing multiple diffusion trajectories by mixing the estimated scores to artificially correlate the generation processes. However, existing methods rely on naive heuristics, such as averaging, without considering task specificity. These approaches do not clarify why such methods work and often produce suboptimal results when a heuristic suitable for one task is blindly applied to others. In this paper, we present a probabilistic framework for analyzing why diffusion synchronization works and reveal where heuristics should be focused; modeling correlations between multiple trajectories and adapting them to each specific task. We further identify optimal correlation models per task, achieving better results than previous approaches that apply a single heuristic across all tasks without justification.
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。