---
title: "DANCE — paper by Zou et al. (2026) — summary and ideas"
description: "The work matters because it offers cheap, backpropagation-free adaptation for real-time deployment. 47 references."
authors: ["Shunge Zou", "Changhu Wang", "Wei Ju", "Ziyue Qiao", "Xiao Luo"]
year: 2026
venue: "Findings of the Association for Computational Linguistics: ACL 2026"
source: "https://aclanthology.org/2026.findings-acl.1860/"
doi: "10.18653/v1/2026.findings-acl.1860"
licence: "CC BY 4.0 (ACL Anthology)"
licence_url: "https://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/zou-2026-dance-diversity-attended-dynamic-caching/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/zou-2026-dance-diversity-attended-dynamic-caching/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2026](https://md.research.elamur.ai/free/en/papers/2026/page/29/) / DANCE: Diversity-attended Dynamic Caching with Asymmetric Quantization for Test-time Adaptation of Vision-Language Models

# DANCE: Diversity-attended Dynamic Caching with Asymmetric Quantization for Test-time Adaptation of Vision-Language Models

**Authors:** Shunge Zou, Changhu Wang, [Wei Ju](https://md.research.elamur.ai/free/en/authors/wei-ju/), [Ziyue Qiao](https://md.research.elamur.ai/free/en/authors/ziyue-qiao/) and [Xiao Luo](https://md.research.elamur.ai/free/en/authors/xiao-luo/) · **Year:** 2026 · **Venue:** [Findings of the Association for Computational Linguistics: ACL 2026](https://md.research.elamur.ai/free/en/venues/findings/) ([Findings of the Association for Computational Linguistics 2026](https://md.research.elamur.ai/free/en/venues/findings/2026/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2026.findings-acl.1860/) · [PDF](https://aclanthology.org/2026.findings-acl.1860.pdf) · [DOI 10.18653/v1/2026.findings-acl.1860](https://doi.org/10.18653/v1/2026.findings-acl.1860) · [OpenAlex](https://openalex.org/W7166857255)

**Citations:** 0 citations (OpenAlex, 2026-10-14)

## Summary
The paper addresses test-time adaptation (TTA) of vision-language models such as CLIP, where only unlabeled test data arrive as a stream. Prior cache-based TTA methods, such as DPE, keep low-entropy samples in memory. According to the authors, this fills the cache with near-identical easy examples, narrows its coverage of the shifting distribution, and lets class prototypes drift. They also argue that cosine retrieval between CLIP image features separates classes poorly.

DANCE changes both how the cache is maintained and how it is queried:
- **Stable view fusion:** it combines augmented views of each image by checking their consistency, not only their confidence.
- **Cache admission:** it scores candidate samples by both reliability (low entropy) and novelty, and evicts the lowest-scoring entries.
- **Asymmetric quantization (AQ):** cached features are stored as binary codes, while queries stay real-valued. This similarity is used both for updating the cache and for correcting logits.

Experiments on ImageNet and 10 cross-domain datasets show gains over representative TTA methods. Diagnostics indicate a more accurate cache and more reliable retrieval. The work matters because it offers cheap, backpropagation-free adaptation for real-time deployment.

## Key points
- Novelty is measured as one minus the highest similarity to cached entries. A single novelty weight, λ_nov = 0.2, is used in all experiments, and the sensitivity analysis shows a broad stable range.
- Replacing AQ with cosine retrieval yields the following on DTD:
  - Retr@1: 47.64% → 52.66%
  - HurtHit: 1.07% → 0.62%
  - top-1 accuracy: 50.16% → 55.12%
- AQ also gives consistent, smaller improvements on Flowers, Caltech101 and Cars. On Cars, for example, HurtHit falls from 2.21% to 1.03%.
- In ablations, removing any single component (view selection, diversity-attended admission or AQ retrieval) lowers accuracy across backbones. Removing combinations of components hurts more.
- On DTD, DANCE keeps cache prototypes closer to oracle class means than DPE does. It also achieves higher cache accuracy on four cross-dataset targets.
- The authors note a limitation: the method does not handle a backbone that is overconfident on out-of-distribution noise, which could cause prototype drift under strong adversarial perturbations.

*Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.*

## Institutions

- [Peking University](https://md.research.elamur.ai/free/en/universities/peking-university/) — University, Beijing, China · 315 works · 43,267 citations
- [University of Wisconsin–Madison](https://md.research.elamur.ai/free/en/universities/university-of-wisconsin-madison/) — University, Madison, United States · 54 works · 49,089 citations
- [Great Bay University](https://md.research.elamur.ai/free/en/universities/great-bay-university/) — University, Dongguan, China · 4 works · 3 citations

Also on the byline: [Fred Hutch Cancer Center](https://ror.org/007ps6h72).

📚 Shelves (3): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 18,614 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 17,965 works · [👁️ Computer Vision and Pattern Recognition](https://md.research.elamur.ai/free/en/shelves/computer-science/computer-vision-and-pattern-recognition/) · 3,486 works

🧭 Topics (3): [🧠 Language models & pretraining](https://md.research.elamur.ai/free/en/topics/language-models/) · 3,894 works · [🖼️ Vision & language](https://md.research.elamur.ai/free/en/topics/multimodal/) · 811 works · [🪶 Efficient models & compression](https://md.research.elamur.ai/free/en/topics/efficient-models/) · 297 works

## Cite

Cite as: Shunge Zou, Changhu Wang, Wei Ju, Ziyue Qiao and Xiao Luo. 2026. DANCE: Diversity-attended Dynamic Caching with Asymmetric Quantization for Test-time Adaptation of Vision-Language Models. Findings of the Association for Computational Linguistics: ACL 2026.

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@inproceedings{zou2026dance,
  title = {{DANCE: Diversity-attended Dynamic Caching with Asymmetric Quantization for Test-time Adaptation of Vision-Language Models}},
  author = {Shunge Zou, Changhu Wang, Wei Ju, Ziyue Qiao and Xiao Luo},
  booktitle = {Findings of the Association for Computational Linguistics: ACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-acl.1860},
  url = {https://doi.org/10.18653/v1/2026.findings-acl.1860}
}
```

- [← previous DaMo: Data Mixing Optimizer in Fine-tuning Multimodal LLMs for Mobile Phone Agents](https://md.research.elamur.ai/free/en/papers/shi-2026-damo-data-mixing-optimizer-fine/)
- [next → Danger Depends on the Mind: A Theory-of-Mind Grounded Dataset and Model for Context-Dependent Dangerous Speech](https://md.research.elamur.ai/free/en/papers/shi-2026-danger-depends-mind-theory-mind/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/29/)

## References

47 references, 5 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Lukas Bossard, Matthieu Guillaumin, and Luc Van Gool. 2014. Food-101–mining discriminative components with random forests. In European conference on computer vision, pages 446–461. Springer. [find](https://scholar.google.com/scholar?q=Lukas+Bossard%2C+Matthieu+Guillaumin%2C+and+Luc+Van+Gool%2E+2014%2E+Food%2D101%E2%80%93mining+discriminative+components+with+random+forests%2E+In+European+conference+on+computer+vision%2C+pages+446%E2%80%93461%2E+Springer%2E)
2. Xinqi Fan, Xueli Chen, Luoxiao Yang, Chuin Hong Yap, Rizwan Qureshi, Qi Dou, Moi Hoon Yap, and Mubarak Shah. 2025. Test-time retrieval-augmented adaptation for vision-language models. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 8810–8819. [find](https://scholar.google.com/scholar?q=Xinqi+Fan%2C+Xueli+Chen%2C+Luoxiao+Yang%2C+Chuin+Hong+Yap%2C+Rizwan+Qureshi%2C+Qi+Dou%2C+Moi+Hoon+Yap%2C+and+Mubarak+Shah%2E+2025%2E+Test%2Dtime+retrieval%2Daugmented+adaptation+for+vision%2Dlanguage+models%2E+In+Proceedings+o)
3. Li Fei-Fei, Rob Fergus, and Pietro Perona. 2004. Learning generative visual models from few training examples: An incremental bayesian approach tested on 101 object categories. In 2004 conference on computer vision and pattern recognition workshop, pages 178–178. IEEE. [find](https://scholar.google.com/scholar?q=Li+Fei%2DFei%2C+Rob+Fergus%2C+and+Pietro+Perona%2E+2004%2E+Learning+generative+visual+models+from+few+training+examples%3A+An+incremental+bayesian+approach+tested+on+101+object+categories%2E+In+2004+conference+on+c)
4. Chun-Mei Feng, Kai Yu, Yong Liu, Salman Khan, and Wangmeng Zuo. 2023. Diverse data augmentation with diffusions for effective test-time prompt tuning. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 2704–2714. [find](https://scholar.google.com/scholar?q=Chun%2DMei+Feng%2C+Kai+Yu%2C+Yong+Liu%2C+Salman+Khan%2C+and+Wangmeng+Zuo%2E+2023%2E+Diverse+data+augmentation+with+diffusions+for+effective+test%2Dtime+prompt+tuning%2E+In+Proceedings+of+the+IEEE%2FCVF+International+Conf)
5. Zenghao Guan, Zhou Yucan, Wu Liu, and Xiaoyan Gu. Statistics caching test-time adaptation for vision-language models. In The Thirty-ninth Annual Conference on Neural Information Processing Systems. [find](https://scholar.google.com/scholar?q=Zenghao+Guan%2C+Zhou+Yucan%2C+Wu+Liu%2C+and+Xiaoyan+Gu%2E+Statistics+caching+test%2Dtime+adaptation+for+vision%2Dlanguage+models%2E+In+The+Thirty%2Dninth+Annual+Conference+on+Neural+Information+Processing+Systems%2E)
6. Patrick Helber, Benjamin Bischke, Andreas Dengel, and Damian Borth. 2019. Eurosat: A novel dataset and deep learning benchmark for land use and land cover classification. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 12(7):2217–2226. [find](https://scholar.google.com/scholar?q=Patrick+Helber%2C+Benjamin+Bischke%2C+Andreas+Dengel%2C+and+Damian+Borth%2E+2019%2E+Eurosat%3A+A+novel+dataset+and+deep+learning+benchmark+for+land+use+and+land+cover+classification%2E+IEEE+Journal+of+Selected+Topi)
7. Dan Hendrycks, Steven Basart, Norman Mu, Saurav Kadavath, Frank Wang, Evan Dorundo, Rahul Desai, Tyler Zhu, Samyak Parajuli, Mike Guo, and 1 others. 2021a. The many faces of robustness: A critical analysis of out-of-distribution generalization. In Proceedings of the IEEE/CVF international conference… [find](https://scholar.google.com/scholar?q=Dan+Hendrycks%2C+Steven+Basart%2C+Norman+Mu%2C+Saurav+Kadavath%2C+Frank+Wang%2C+Evan+Dorundo%2C+Rahul+Desai%2C+Tyler+Zhu%2C+Samyak+Parajuli%2C+Mike+Guo%2C+and+1+others%2E+2021a%2E+The+many+faces+of+robustness%3A+A+critical+ana)
8. Dan Hendrycks, Kevin Zhao, Steven Basart, Jacob Steinhardt, and Dawn Song. 2021b. Natural adversarial examples. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 15262–15271. [find](https://scholar.google.com/scholar?q=Dan+Hendrycks%2C+Kevin+Zhao%2C+Steven+Basart%2C+Jacob+Steinhardt%2C+and+Dawn+Song%2E+2021b%2E+Natural+adversarial+examples%2E+In+Proceedings+of+the+IEEE%2FCVF+conference+on+computer+vision+and+pattern+recognition%2C+pa)
9. Zhaohong Huang, Yuxin Zhang, Jingjing Xie, Fei Chao, and Rongrong Ji. 2025. Gs-bias: Global-spatial bias learner for single-image test-time adaptation of vision-language models. arXiv preprint arXiv:2507.11969. — [arXiv:2507.11969](https://arxiv.org/abs/2507.11969) [find](https://scholar.google.com/scholar?q=Zhaohong+Huang%2C+Yuxin+Zhang%2C+Jingjing+Xie%2C+Fei+Chao%2C+and+Rongrong+Ji%2E+2025%2E+Gs%2Dbias%3A+Global%2Dspatial+bias+learner+for+single%2Dimage+test%2Dtime+adaptation+of+vision%2Dlanguage+models%2E+arXiv+preprint+arXiv%3A2)
10. Chao Jia, Yinfei Yang, Ye Xia, Yi-Ting Chen, Zarana Parekh, Hieu Pham, Quoc Le, Yun-Hsuan Sung, Zhen Li, and Tom Duerig. 2021. [Scaling up visual and vision-language representation learning with noisy text supervision](https://md.research.elamur.ai/free/en/papers/jia-2021-scaling-up-visual-vision-language/). In International conference on machine learning, pages 4904–4916. PMLR.
11. Siddharth Joshi, Arnav Jain, Ali Payani, and Baharan Mirzasoleiman. 2024. Data-efficient contrastive language-image pretraining: Prioritizing data quality over quantity. In International Conference on Artificial Intelligence and Statistics, pages 1000–1008. PMLR. [find](https://scholar.google.com/scholar?q=Siddharth+Joshi%2C+Arnav+Jain%2C+Ali+Payani%2C+and+Baharan+Mirzasoleiman%2E+2024%2E+Data%2Defficient+contrastive+language%2Dimage+pretraining%3A+Prioritizing+data+quality+over+quantity%2E+In+International+Conference+on)
12. Adilbek Karmanov, Dayan Guan, Shijian Lu, Abdulmotaleb El Saddik, and Eric Xing. 2024. Efficient test-time adaptation of vision-language models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 14162– 14171. [find](https://scholar.google.com/scholar?q=Adilbek+Karmanov%2C+Dayan+Guan%2C+Shijian+Lu%2C+Abdulmotaleb+El+Saddik%2C+and+Eric+Xing%2E+2024%2E+Efficient+test%2Dtime+adaptation+of+vision%2Dlanguage+models%2E+In+Proceedings+of+the+IEEE%2FCVF+Conference+on+Computer+V)
13. Jonathan Krause, Michael Stark, Jia Deng, and Li Fei-Fei. 2013. 3d object representations for fine-grained categorization. In Proceedings of the IEEE international conference on computer vision workshops, pages 554–561. [find](https://scholar.google.com/scholar?q=Jonathan+Krause%2C+Michael+Stark%2C+Jia+Deng%2C+and+Li+Fei%2DFei%2E+2013%2E+3d+object+representations+for+fine%2Dgrained+categorization%2E+In+Proceedings+of+the+IEEE+international+conference+on+computer+vision+worksh)
14. Shuaifeng Li, Mao Ye, Lihua Zhou, Nianxin Li, Siying Xiao, Song Tang, and Xiatian Zhu. 2024. Cloud object detector adaptation by integrating different source knowledge. Advances in Neural Information Processing Systems, 37:25251–25283. [find](https://scholar.google.com/scholar?q=Shuaifeng+Li%2C+Mao+Ye%2C+Lihua+Zhou%2C+Nianxin+Li%2C+Siying+Xiao%2C+Song+Tang%2C+and+Xiatian+Zhu%2E+2024%2E+Cloud+object+detector+adaptation+by+integrating+different+source+knowledge%2E+Advances+in+Neural+Information+)
15. Yiwen Liang, Hui Chen, Yizhe Xiong, Zihan Zhou, Mengyao Lyu, Zijia Lin, Shuaicheng Niu, Sicheng Zhao, Jungong Han, and Guiguang Ding. 2025b. Advancing reliable test-time adaptation of vision-language models under visual variations. In Proceedings of the 33rd ACM International Conference on Multimedi… [find](https://scholar.google.com/scholar?q=Yiwen+Liang%2C+Hui+Chen%2C+Yizhe+Xiong%2C+Zihan+Zhou%2C+Mengyao+Lyu%2C+Zijia+Lin%2C+Shuaicheng+Niu%2C+Sicheng+Zhao%2C+Jungong+Han%2C+and+Guiguang+Ding%2E+2025b%2E+Advancing+reliable+test%2Dtime+adaptation+of+vision%2Dlanguage+)
16. Yu-An Liu, Ruqing Zhang, Jiafeng Guo, Maarten de Rijke, Yixing Fan, and Xueqi Cheng. 2025. Robust neural information retrieval: An adversarial and out-of-distribution perspective. ACM Transactions on Information Systems, 44(1):1–48. [find](https://scholar.google.com/scholar?q=Yu%2DAn+Liu%2C+Ruqing+Zhang%2C+Jiafeng+Guo%2C+Maarten+de+Rijke%2C+Yixing+Fan%2C+and+Xueqi+Cheng%2E+2025%2E+Robust+neural+information+retrieval%3A+An+adversarial+and+out%2Dof%2Ddistribution+perspective%2E+ACM+Transactions+on+)
17. Junyu Luo, Xiao Luo, Xiusi Chen, Zhiping Xiao, Wei Ju, and Ming Zhang. 2025a. Semi-supervised fine-tuning for large language models. In Findings of the Association for Computational Linguistics: NAACL 2025, pages 2795–2808. [find](https://scholar.google.com/scholar?q=Junyu+Luo%2C+Xiao+Luo%2C+Xiusi+Chen%2C+Zhiping+Xiao%2C+Wei+Ju%2C+and+Ming+Zhang%2E+2025a%2E+Semi%2Dsupervised+fine%2Dtuning+for+large+language+models%2E+In+Findings+of+the+Association+for+Computational+Linguistics%3A+NAACL)
18. Junyu Luo, Bohan Wu, Xiao Luo, Zhiping Xiao, Yiqiao Jin, Rong-Cheng Tu, Nan Yin, Yifan Wang, Jingyang Yuan, Wei Ju, and 1 others. 2025b. [A survey on efficient large language model training: From data-centric perspectives](https://md.research.elamur.ai/free/en/papers/luo-2025-survey-efficient-large-language-model/). In Proceedings of the 63rd Annual Meeting of the Association for Computational…
19. Subhransu Maji, Esa Rahtu, Juho Kannala, Matthew Blaschko, and Andrea Vedaldi. 2013. Fine-grained visual classification of aircraft. arXiv preprint arXiv:1306.5151. — [arXiv:1306.5151](https://arxiv.org/abs/1306.5151) [find](https://scholar.google.com/scholar?q=Subhransu+Maji%2C+Esa+Rahtu%2C+Juho+Kannala%2C+Matthew+Blaschko%2C+and+Andrea+Vedaldi%2E+2013%2E+Fine%2Dgrained+visual+classification+of+aircraft%2E+arXiv+preprint+arXiv%3A1306%2E5151%2E+%E2%80%94+%5BarXiv%3A1306%2E5151%5D%28https%3A%2F%2Farxiv%2Eo)
20. Yu A Malkov and Dmitry A Yashunin. 2018. [Efficient and robust approximate nearest neighbor search using hierarchical navigable small world graphs](https://md.research.elamur.ai/free/en/papers/malkov-2016-efficient-robust-approximate-nearest-neighbor/). IEEE transactions on pattern analysis and machine intelligence, 42(4):824–836.
21. Behnam Neyshabur, Nati Srebro, Russ R Salakhutdinov, Yury Makarychev, and Payman Yadollahpour. 2013. The power of asymmetry in binary hashing. Advances in neural information processing systems, 26. [find](https://scholar.google.com/scholar?q=Behnam+Neyshabur%2C+Nati+Srebro%2C+Russ+R+Salakhutdinov%2C+Yury+Makarychev%2C+and+Payman+Yadollahpour%2E+2013%2E+The+power+of+asymmetry+in+binary+hashing%2E+Advances+in+neural+information+processing+systems%2C+26%2E)
22. Khanh-Binh Nguyen, Phuoc-Nguyen Bui, Hyunseung Choo, and Duc Thanh Nguyen. 2025. Adaptive cache enhancement for test-time adaptation of vision-language models. arXiv preprint arXiv:2508.07570. — [arXiv:2508.07570](https://arxiv.org/abs/2508.07570) [find](https://scholar.google.com/scholar?q=Khanh%2DBinh+Nguyen%2C+Phuoc%2DNguyen+Bui%2C+Hyunseung+Choo%2C+and+Duc+Thanh+Nguyen%2E+2025%2E+Adaptive+cache+enhancement+for+test%2Dtime+adaptation+of+vision%2Dlanguage+models%2E+arXiv+preprint+arXiv%3A2508%2E07570%2E+%E2%80%94+%5BarXi)
23. Shuaifeng Li, Mao Ye, Xiatian Zhu, Lihua Zhou, and Lin Xiong. 2022. Source-free object detection by learning to overlook domain style. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 8014–8023. [find](https://scholar.google.com/scholar?q=Shuaifeng+Li%2C+Mao+Ye%2C+Xiatian+Zhu%2C+Lihua+Zhou%2C+and+Lin+Xiong%2E+2022%2E+Source%2Dfree+object+detection+by+learning+to+overlook+domain+style%2E+In+Proceedings+of+the+IEEE%2FCVF+conference+on+computer+vision+and+)
24. Maria-Elena Nilsback and Andrew Zisserman. 2008. Automated flower classification over a large number of classes. In 2008 Sixth Indian conference on computer vision, graphics & image processing, pages 722–729. IEEE. [find](https://scholar.google.com/scholar?q=Maria%2DElena+Nilsback+and+Andrew+Zisserman%2E+2008%2E+Automated+flower+classification+over+a+large+number+of+classes%2E+In+2008+Sixth+Indian+conference+on+computer+vision%2C+graphics+%26+image+processing%2C+pages+)
25. Jian Liang, Ran He, and Tieniu Tan. 2025a. A comprehensive survey on test-time adaptation under distribution shifts. International Journal of Computer Vision, 133(1):31–64. [find](https://scholar.google.com/scholar?q=Jian+Liang%2C+Ran+He%2C+and+Tieniu+Tan%2E+2025a%2E+A+comprehensive+survey+on+test%2Dtime+adaptation+under+distribution+shifts%2E+International+Journal+of+Computer+Vision%2C+133%281%29%3A31%E2%80%9364%2E)
26. Omkar M Parkhi, Andrea Vedaldi, Andrew Zisserman, and CV Jawahar. 2012. Cats and dogs. In 2012 IEEE conference on computer vision and pattern recognition, pages 3498–3505. IEEE. [find](https://scholar.google.com/scholar?q=Omkar+M+Parkhi%2C+Andrea+Vedaldi%2C+Andrew+Zisserman%2C+and+CV+Jawahar%2E+2012%2E+Cats+and+dogs%2E+In+2012+IEEE+conference+on+computer+vision+and+pattern+recognition%2C+pages+3498%E2%80%933505%2E+IEEE%2E)
27. Victor Weixin Liang, Yuhui Zhang, Yongchan Kwon, Serena Yeung, and James Y Zou. 2022. Mind the gap: Understanding the modality gap in multi-modal contrastive representation learning. Advances in Neural Information Processing Systems, 35:17612–17625. [find](https://scholar.google.com/scholar?q=Victor+Weixin+Liang%2C+Yuhui+Zhang%2C+Yongchan+Kwon%2C+Serena+Yeung%2C+and+James+Y+Zou%2E+2022%2E+Mind+the+gap%3A+Understanding+the+modality+gap+in+multi%2Dmodal+contrastive+representation+learning%2E+Advances+in+Neura)
28. Sarah Pratt, Ian Covert, Rosanne Liu, and Ali Farhadi. 2023. What does a platypus look like? generating customized prompts for zero-shot image classification. In Proceedings of the IEEE/CVF international conference on computer vision, pages 15691–15701. [find](https://scholar.google.com/scholar?q=Sarah+Pratt%2C+Ian+Covert%2C+Rosanne+Liu%2C+and+Ali+Farhadi%2E+2023%2E+What+does+a+platypus+look+like%3F+generating+customized+prompts+for+zero%2Dshot+image+classification%2E+In+Proceedings+of+the+IEEE%2FCVF+internatio)
29. Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, and 1 others. 2021. [Learning transferable visual models from natural language supervision](https://md.research.elamur.ai/free/en/papers/radford-2021-learning-transferable-visual-models-natural/). In International conference on machine learning, pages 8748–8…
30. Benjamin Recht, Rebecca Roelofs, Ludwig Schmidt, and Vaishaal Shankar. 2019. Do imagenet classifiers generalize to imagenet? In International conference on machine learning, pages 5389–5400. PMLR. [find](https://scholar.google.com/scholar?q=Benjamin+Recht%2C+Rebecca+Roelofs%2C+Ludwig+Schmidt%2C+and+Vaishaal+Shankar%2E+2019%2E+Do+imagenet+classifiers+generalize+to+imagenet%3F+In+International+conference+on+machine+learning%2C+pages+5389%E2%80%935400%2E+PMLR%2E)
31. Xiaohua Zhai, Xiao Wang, Basil Mustafa, Andreas Steiner, Daniel Keysers, Alexander Kolesnikov, and Lucas Beyer. 2022. Lit: Zero-shot transfer with locked-image text tuning. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 18123–18133. [find](https://scholar.google.com/scholar?q=Xiaohua+Zhai%2C+Xiao+Wang%2C+Basil+Mustafa%2C+Andreas+Steiner%2C+Daniel+Keysers%2C+Alexander+Kolesnikov%2C+and+Lucas+Beyer%2E+2022%2E+Lit%3A+Zero%2Dshot+transfer+with+locked%2Dimage+text+tuning%2E+In+Proceedings+of+the+IEEE%2F)
32. Ce Zhang, Simon Stepputtis, Katia Sycara, and Yaqi Xie. 2024a. Dual prototype evolving for test-time generalization of vision-language models. Advances in Neural Information Processing Systems, 37:32111– 32136. [find](https://scholar.google.com/scholar?q=Ce+Zhang%2C+Simon+Stepputtis%2C+Katia+Sycara%2C+and+Yaqi+Xie%2E+2024a%2E+Dual+prototype+evolving+for+test%2Dtime+generalization+of+vision%2Dlanguage+models%2E+Advances+in+Neural+Information+Processing+Systems%2C+37%3A321)
33. Fumin Shen, Yang Yang, Li Liu, Wei Liu, Dacheng Tao, and Heng Tao Shen. 2017. Asymmetric binary coding for image search. IEEE Transactions on Multimedia, 19(9):2022–2032. [find](https://scholar.google.com/scholar?q=Fumin+Shen%2C+Yang+Yang%2C+Li+Liu%2C+Wei+Liu%2C+Dacheng+Tao%2C+and+Heng+Tao+Shen%2E+2017%2E+Asymmetric+binary+coding+for+image+search%2E+IEEE+Transactions+on+Multimedia%2C+19%289%29%3A2022%E2%80%932032%2E)
34. Renrui Zhang, Wei Zhang, Rongyao Fang, Peng Gao, Kunchang Li, Jifeng Dai, Yu Qiao, and Hongsheng Li. 2022. Tip-adapter: Training-free adaption of clip for few-shot classification. In European conference on computer vision, pages 493–510. Springer. [find](https://scholar.google.com/scholar?q=Renrui+Zhang%2C+Wei+Zhang%2C+Rongyao+Fang%2C+Peng+Gao%2C+Kunchang+Li%2C+Jifeng+Dai%2C+Yu+Qiao%2C+and+Hongsheng+Li%2E+2022%2E+Tip%2Dadapter%3A+Training%2Dfree+adaption+of+clip+for+few%2Dshot+classification%2E+In+European+conferen)
35. Manli Shu, Weili Nie, De-An Huang, Zhiding Yu, Tom Goldstein, Anima Anandkumar, and Chaowei Xiao. 2022. Test-time prompt tuning for zero-shot generalization in vision-language models. Advances in Neural Information Processing Systems, 35:14274– 14289. [find](https://scholar.google.com/scholar?q=Manli+Shu%2C+Weili+Nie%2C+De%2DAn+Huang%2C+Zhiding+Yu%2C+Tom+Goldstein%2C+Anima+Anandkumar%2C+and+Chaowei+Xiao%2E+2022%2E+Test%2Dtime+prompt+tuning+for+zero%2Dshot+generalization+in+vision%2Dlanguage+models%2E+Advances+in+Neur)
36. Taolin Zhang, Jinpeng Wang, Hang Guo, Tao Dai, Bin Chen, and Shu-Tao Xia. 2024b. Boostadapter: Improving vision-language test-time adaptation via regional bootstrapping. Advances in Neural Information Processing Systems, 37:67795–67825. [find](https://scholar.google.com/scholar?q=Taolin+Zhang%2C+Jinpeng+Wang%2C+Hang+Guo%2C+Tao+Dai%2C+Bin+Chen%2C+and+Shu%2DTao+Xia%2E+2024b%2E+Boostadapter%3A+Improving+vision%2Dlanguage+test%2Dtime+adaptation+via+regional+bootstrapping%2E+Advances+in+Neural+Information)
37. Khurram Soomro, Amir Roshan Zamir, and Mubarak Shah. 2012. A dataset of 101 human action classes from videos in the wild. Center for Research in Computer Vision, 2(11):1–7. [find](https://scholar.google.com/scholar?q=Khurram+Soomro%2C+Amir+Roshan+Zamir%2C+and+Mubarak+Shah%2E+2012%2E+A+dataset+of+101+human+action+classes+from+videos+in+the+wild%2E+Center+for+Research+in+Computer+Vision%2C+2%2811%29%3A1%E2%80%937%2E)
38. Kaiyang Zhou, Jingkang Yang, Chen Change Loy, and Ziwei Liu. 2022a. Conditional prompt learning for vision-language models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 16816–16825. [find](https://scholar.google.com/scholar?q=Kaiyang+Zhou%2C+Jingkang+Yang%2C+Chen+Change+Loy%2C+and+Ziwei+Liu%2E+2022a%2E+Conditional+prompt+learning+for+vision%2Dlanguage+models%2E+In+Proceedings+of+the+IEEE%2FCVF+conference+on+computer+vision+and+pattern+rec)
39. Haohan Wang, Songwei Ge, Zachary Lipton, and Eric P Xing. 2019. Learning robust global representations by penalizing local predictive power. Advances in neural information processing systems, 32. [find](https://scholar.google.com/scholar?q=Haohan+Wang%2C+Songwei+Ge%2C+Zachary+Lipton%2C+and+Eric+P+Xing%2E+2019%2E+Learning+robust+global+representations+by+penalizing+local+predictive+power%2E+Advances+in+neural+information+processing+systems%2C+32%2E)
40. Kaiyang Zhou, Jingkang Yang, Chen Change Loy, and Ziwei Liu. 2022b. Learning to prompt for vision-language models. International Journal of Computer Vision, 130(9):2337–2348. [find](https://scholar.google.com/scholar?q=Kaiyang+Zhou%2C+Jingkang+Yang%2C+Chen+Change+Loy%2C+and+Ziwei+Liu%2E+2022b%2E+Learning+to+prompt+for+vision%2Dlanguage+models%2E+International+Journal+of+Computer+Vision%2C+130%289%29%3A2337%E2%80%932348%2E)
41. Jianxiong Xiao, James Hays, Krista A Ehinger, Aude Oliva, and Antonio Torralba. 2010. Sun database: Large-scale scene recognition from abbey to zoo. In 2010 IEEE computer society conference on computer vision and pattern recognition, pages 3485–3492. IEEE. [find](https://scholar.google.com/scholar?q=Jianxiong+Xiao%2C+James+Hays%2C+Krista+A+Ehinger%2C+Aude+Oliva%2C+and+Antonio+Torralba%2E+2010%2E+Sun+database%3A+Large%2Dscale+scene+recognition+from+abbey+to+zoo%2E+In+2010+IEEE+computer+society+conference+on+compute)
42. Lihua Zhou, Mao Ye, Shuaifeng Li, Nianxin Li, Xiatian Zhu, Lei Deng, Hongbin Liu, and Zhen Lei. 2025a. Bayesian test-time adaptation for vision-language models. In Proceedings of the Computer Vision and Pattern Recognition Conference, pages 29999–30009. [find](https://scholar.google.com/scholar?q=Lihua+Zhou%2C+Mao+Ye%2C+Shuaifeng+Li%2C+Nianxin+Li%2C+Xiatian+Zhu%2C+Lei+Deng%2C+Hongbin+Liu%2C+and+Zhen+Lei%2E+2025a%2E+Bayesian+test%2Dtime+adaptation+for+vision%2Dlanguage+models%2E+In+Proceedings+of+the+Computer+Vision+a)
43. Yizhe Xiong, Hui Chen, Zijia Lin, Sicheng Zhao, and Guiguang Ding. 2023. Confidence-based visual dispersal for few-shot unsupervised domain adaptation. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 11621–11631. [find](https://scholar.google.com/scholar?q=Yizhe+Xiong%2C+Hui+Chen%2C+Zijia+Lin%2C+Sicheng+Zhao%2C+and+Guiguang+Ding%2E+2023%2E+Confidence%2Dbased+visual+dispersal+for+few%2Dshot+unsupervised+domain+adaptation%2E+In+Proceedings+of+the+IEEE%2FCVF+International+Con)
44. Yujun Zhou, Zhenwen Liang, Haolin Liu, Wenhao Yu, Kishan Panaganti, Linfeng Song, Dian Yu, Xiangliang Zhang, Haitao Mi, and Dong Yu. 2025b. Evolving language models without labels: Majority drives selection, novelty promotes variation. arXiv preprint arXiv:2509.15194. — [arXiv:2509.15194](https://arxiv.org/abs/2509.15194) [find](https://scholar.google.com/scholar?q=Yujun+Zhou%2C+Zhenwen+Liang%2C+Haolin+Liu%2C+Wenhao+Yu%2C+Kishan+Panaganti%2C+Linfeng+Song%2C+Dian+Yu%2C+Xiangliang+Zhang%2C+Haitao+Mi%2C+and+Dong+Yu%2E+2025b%2E+Evolving+language+models+without+labels%3A+Majority+drives+sel)
45. Jiahui Yu, Zirui Wang, Vijay Vasudevan, Legg Yeung, Mojtaba Seyedhosseini, and Yonghui Wu. 2022. [Coca: Contrastive captioners are image-text foundation models](https://md.research.elamur.ai/free/en/papers/yu-2022-coca-contrastive-captioners-image-text/). arXiv preprint arXiv:2205.01917.
46. Tao Yu, Zhihe Lu, Xin Jin, Zhibo Chen, and Xinchao Wang. 2023. Task residual for tuning vision-language models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 10899–10909. [find](https://scholar.google.com/scholar?q=Tao+Yu%2C+Zhihe+Lu%2C+Xin+Jin%2C+Zhibo+Chen%2C+and+Xinchao+Wang%2E+2023%2E+Task+residual+for+tuning+vision%2Dlanguage+models%2E+In+Proceedings+of+the+IEEE%2FCVF+conference+on+computer+vision+and+pattern+recognition%2C+pa)
47. Haotian Zhai, Xinyu Chen, Can Zhang, Tianming Sha, and Ruirui Li. 2025. Mitigating cache noise in test-time adaptation for large vision-language models. In 2025 IEEE International Conference on Multimedia and Expo (ICME), pages 1–6. IEEE. [find](https://scholar.google.com/scholar?q=Haotian+Zhai%2C+Xinyu+Chen%2C+Can+Zhang%2C+Tianming+Sha%2C+and+Ruirui+Li%2E+2025%2E+Mitigating+cache+noise+in+test%2Dtime+adaptation+for+large+vision%2Dlanguage+models%2E+In+2025+IEEE+International+Conference+on+Multim)

*Updated 2 November 2026*