重新思考策略性扩散蒸馏中的无分类器引导

中文摘要 AI 辅助

策略性蒸馏（OPD）通过在当前学生模型生成的轨迹上查询教师模型来调整扩散模型，但对于现代扩散系统的默认组件无分类器引导（CFG）下它应如何表现，人们仍知之甚少。现有OPD方法自然地将速度匹配扩展到由CFG组成的预测，直接匹配教师和学生的引导速度。我们表明，此目标在分支级别未得到充分识别：正分支和负分支误差可在引导预测中相互补偿。通过两个对比案例，我们发现，在共享负条件下，简单匹配仍然有效，此时两个分支误差会共同减小。然而，当模型的原生CFG模式在教师的负分支中保留了学生无法获得的特权信息时，这种共同减小就会失效，并且组合目标会引发对抗性分支误差动态，减小正分支误差同时增加负分支误差。我们将这种失败模式称为负分支不对称（NBA）。为解决NBA，我们引入正向匹配（PDM），这是一种分支感知的OPD目标，它分别约束正预测和CFG条件方向。我们将PDM应用于从密集到稀疏的视频控制，其中简单的引导匹配对推理引导尺度高度敏感，而分支感知监督能够实现更强大且有效的知识转移。

英文摘要

On-policy distillation (OPD) adapts diffusion models by querying a teacher along trajectories generated by the current student, but how it should behave under classifier-free guidance (CFG), a default component of modern diffusion systems, remains poorly understood. Existing OPD methods naturally extend velocity matching to the CFG-composed prediction, directly matching teacher and student guided velocities. We show that this objective is under-identified at the branch level: positive- and negative-branch errors can compensate in the guided prediction. Through two contrasting cases, we find that naive matching remains effective under shared negative conditioning, where both branch errors decrease jointly. When the model's native CFG schema retains privileged information in the teacher's negative branch that is unavailable to the student, however, this joint reduction breaks down and the composed objective induces antagonistic branch-error dynamics, reducing the positive-branch error while increasing the negative-branch error. We term this failure mode Negative Branch Asymmetry (NBA). To address NBA, we introduce Positive--Direction Matching (PDM), a branch-aware OPD objective that separately constrains the positive prediction and the CFG conditional direction. We apply PDM to dense-to-sparse video control, where naive guided matching is highly sensitive to inference guidance scales, while branch-aware supervision enables more robust and effective knowledge transfer.

发表机构

• Qwen Business Unit of Alibaba（阿里巴巴的通义千问业务部）

机构由 AI 辅助整理，请以论文原文为准。