The Professor: Multi-Teacher Unsupervised Prompt Distillation for Vision-Language Models
Quick Answer
TheProfessor introduces a multi-teacher approach for prompt distillation in vision-language models, enhancing performance on datasets like EuroSAT, achieving a +5.78 HM improvement.
Quick Take
By leveraging a two-teacher ensemble, it outperforms single-teacher methods, demonstrating significant gains in domain-shifted scenarios.
Key Points
- TheProfessor distills from a two-teacher ensemble: a domain-finetuned ViT-L/14 and a zero-shot EVA-CLIP-L/14.
- Confidence-weighted ensembling boosts average HM from 87.52 to 89.28 across datasets.
- Improvements vary by dataset, with EuroSAT showing the largest gain of +5.78 HM.
- Single-teacher PromptKD and equal-probability ensembling were also evaluated for comparison.
- Results indicate multi-teacher distillation is beneficial under domain shifts.
Paper Resources
📖 Reader Mode
~2 min readAbstract: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.
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2606.23897 [cs.CV] |
| (or arXiv:2606.23897v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2606.23897 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Ahmed Alzuhair [view email]
[v1]
Mon, 22 Jun 2026 19:53:41 UTC (1,499 KB)
— Originally published at arxiv.org
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.CV
See more →ProMoE-FL: Prototype-conditioned Mixture of Experts for Multimodal Federated Learning with Missing Modalities
ProMoE-FL introduces a Prototype-conditioned Mixture-of-Experts framework for multimodal federated learning, effectively addressing missing modalities. It outperforms existing methods on four chest X-ray datasets, demonstrating superior feature synthesis capabilities in both homogeneous and heterogeneous settings.