The Professor: 面向视觉语言模型的多教师无监督提示蒸馏
The Professor: Multi-Teacher Unsupervised Prompt Distillation for Vision-Language Models
AI总结：
提出多教师提示蒸馏方法TheProfessor，融合领域微调教师和零样本教师，在四个数据集上平均HM提升1.77点，尤其对域偏移数据集效果显著。
摘要：
提示蒸馏通过匹配教师模型在无标注领域图像上的预测，将大型视觉语言模型（如CLIP）压缩为轻量级学生模型。PromptKD（CVPR 2024）使用单个PromptSRC微调的ViT-L/14教师和ViT-B/16学生建立了这一范式。我们提出TheProfessor，一种多教师扩展方法，从固定的双教师集成中蒸馏：一个领域微调的PromptSRC ViT-L/14教师和一个零样本EVA-CLIP-L/14教师（其logits按数据集预计算）。我们在四个基础到新颖数据集（Caltech-101、DTD、UCF101和EuroSAT）上评估了单教师PromptKD、等概率集成和置信度加权集成。在12次单种子实验中，置信度加权集成将平均HM从87.52提高到89.28（+1.77点），而等概率平均将平均HM提高到88.88（+1.37点）。增益依赖于数据集：在Caltech-101上可忽略（置信度加权HM+0.16），在UCF101上中等（+0.62），在域偏移的EuroSAT上最大（+5.78）。这些结果更新了我们早期仅针对Caltech-101的分析，表明当第二教师在域偏移下提供互补监督时，多教师提示蒸馏最为有用。
英文摘要：
Prompt distillation compresses large vision-language models (VLMs) such as CLIP into lightweight student models by matching teacher predictions on unlabeled domain images. PromptKD (CVPR 2024) established this paradigm with a single PromptSRC-finetuned ViT-L/14 teacher and a ViT-B/16 student. We propose TheProfessor, a multi-teacher extension that distills from a fixed two-teacher ensemble: a domain-finetuned PromptSRC ViT-L/14 teacher and a zero-shot EVA-CLIP-L/14 teacher whose logits are pre-computed per dataset. We evaluate single-teacher PromptKD, equal-probability ensembling, and confidence-weighted ensembling on four base-to-novel datasets: Caltech-101, DTD, UCF101, and EuroSAT. In a 12-run single-seed sweep, confidence-weighted ensembling improves average HM from 87.52 to 89.28 (+1.77 points), while equal averaging improves average HM to 88.88 (+1.37 points). Gains are dataset dependent: they are negligible on Caltech-101 (+0.16 HM for confidence weighting), modest on UCF101 (+0.62), and largest on domain-shifted EuroSAT (+5.78). These results update our earlier Caltech-only analysis and show that multi-teacher prompt distillation is most useful when the second teacher contributes complementary supervision under domain shift.
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。
