净化后对齐：提升扩散模型表示对齐效果
原标题：Purify Before You Align: Improving Representation Alignment for Diffusion Models
AI 摘要
该论文提出局部外观净化（LAP）方法，通过移除预训练视觉编码器表示中可由局部外观预测的内容，仅对齐残差信息，从而提升扩散模型的生成质量。实验在MAE、AIMv2、CLIP、DINOv2、DINOv3五种编码器上验证，其中MAE提升最大，在ImageNet-256上使用SiT-B/2将FID显著降低。作者还给出信息论解释和一种可在训练前计算的常数读出诊断分数，用于预测哪些编码器能从净化中受益。
正文节选
Representation Alignment for Diffusion Models Abstract Representation alignment improves and accelerates diffusion training by encouraging intermediate features to match a pretrained visual encoder. However, standard alignment matches pretrained representations without distinguishing what information may be more useful for generation. We introduce local appearance purification (LAP), which improves generation quality and makes encoders such as MAE more effective alignment targets by removing loc