Pith. sign in

REVIEW 3 major objections 2 minor 81 references

Multi-Cache Enhanced Prototype Learning for Test-Time Generalization of Vision-Language Models

T0 review · 3 major / 2 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read The paper proposes MCP, a test-time adaptation method that builds prototypes from three caches—entropy-selected samples, aligned visual–textual features, and high-entropy negatives—to improve vision-language model generalization under…

desk verdict The submitted full text is an unrelated forest-ecology paper, so there is no MCP paper here to evaluate. read the letter →

arxiv 2508.01225 v2 pith:2DE7UKB6 submitted 2025-08-02 cs.CV cs.AI

Xinyu Chen , Haotian Zhai , Can Zhang , Xiupeng Shi , Ruirui Li This is my paper
classification cs.CVcs.AI
keywords test-timeadaptationvision-languagemodelsprototypelearningcache-basedmethodsdistributionshiftzero-shotgeneralizationintra-classcompactness
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper is trying to establish that test-time adaptation of vision-language models should not rest on low-entropy sample selection alone. It reports a positive correlation between cache-enhanced adaptation performance and intra-class compactness, and uses that correlation to motivate a multi-cache prototype method: one cache initializes prototypes, one aligns visual and textual features for compactness, and one calibrates predictions using high-entropy samples. If the correlation and the method hold up, cache-based adaptation can improve under distribution shift without any labels. The MCP++ variant adds prototype residual fine-tuning for further gains, and across 15 downstream tasks the authors report state-of-the-art generalization.

What carries the argument

The central machinery is a set of three caches, where a cache is a stored bank of features drawn from unlabeled test data and the frozen vision-language model. An entropy cache selects low-entropy samples to initialize class prototypes; an align cache combines visual and textual information so that each class's prototypes form a compact intra-class distribution; a negative cache collects high-entropy samples and uses them to calibrate predictions away from wrong classes. In MCP++, cross-modal prototype alignment and prototype residual fine-tuning refine the prototypes further.

What would settle it

The supplied full text for this review is an unrelated forest-ecology manuscript, so the 15-task tables behind the claim are not visible here; the claim would be settled by running MCP against the entropy-only cache baseline on a corrupted or shifted benchmark and checking whether the align and negative caches add accuracy when low-entropy samples are demonstrably mislabeled.

Watch

Extended reading notes

Core claim

On its own terms, the paper's central claim is that existing cache-enhanced test-time adaptation methods fail because they assume low-entropy samples are reliable under distribution shift, whereas the real driver of performance is intra-class compactness. The paper states this as a discovered correlation: cache-enhanced performance increases as the prototypes produce more compact intra-class distributions. MCP is built to exploit that correlation directly, with an entropy cache for initialization, an align cache for visual–textual integration, and a negative cache that turns high-entropy samples into a calibration signal. The proposed MCP++ framework extends this with cross-modal prototype alignment, residual learning, and prototype residual fine-tuning.

Load-bearing premise

The method assumes that a test sample the frozen model already predicts with low entropy is a reliable prototype anchor, and that compacting classes in feature space transfers to accuracy; if confident predictions are systematically wrong in the new distribution, the cache design loses its foundation.

Editorial extensions

If this is right

  • Discarding high-entropy samples is not enough: they carry calibration signal that can improve predictions under distribution shift.
  • Aligning visual and textual features before prototype construction yields more compact intra-class distributions and better generalization.
  • The method operates in the zero-shot, no-label setting, using only unlabeled test data together with the frozen vision-language model.
  • The MCP++ extension shows that prototype residual fine-tuning can push adaptation further when additional compute is available.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Inference: the negative cache could be reused as a calibration set, so a natural extension is to test whether MCP also reduces expected calibration error, not just top-1 accuracy.
  • Inference: because the caches operate on the model's own embeddings, the same recipe should transfer to other vision-language architectures; rerunning the 15-task suite on a different backbone would test this.
  • Inference: the align cache implicitly addresses the visual–textual modality gap, and separating this correction from prototype construction might connect the method to unsupervised domain adaptation.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 2 minor

Summary. The manuscript as submitted carries the title and abstract of a computer-vision paper on test-time adaptation of vision-language models, proposing a Multi-Cache enhanced Prototype-based Test-Time Adaptation (MCP) method with entropy, align, and negative caches, and claiming state-of-the-art generalization across 15 downstream tasks. However, the supplied full text is an entirely unrelated manuscript, 'Inferring Processes Within Dynamic Forest Models Using Hybrid Modeling' by Maximilian Pichler and Yannek Käber (arXiv:2508.01228v2), which concerns forest dynamics modeling with the FINN hybrid approach. No equations, experiments, ablations, or hyperparameters of the MCP method appear anywhere in the provided record. The only evidence for the MCP claims is the abstract itself, which is insufficient for any technical evaluation.

Significance. Were the MCP method and results present as described, the paper could be significant for the test-time adaptation community: cache-based methods are an active area, and the proposed combination of entropy, align, and negative caches, plus an MCP++ variant with cross-modal alignment and residual fine-tuning, would plausibly offer a useful contribution. The claimed identification of a positive correlation between cache-enhanced performance and intra-class compactness is a testable hypothesis with potential practical value. However, none of these contributions can be evaluated, reproduced, or even inspected from the submitted material. The supplied full text does not mention vision-language models, caches, prototypes, test-time adaptation, or any of the claimed experiments. The significance is therefore entirely prospective and cannot be assessed on the evidence in front of me.

major comments (3)
  1. [Full text (entire)] The full text supplied is not the manuscript referenced by the abstract. It is a paper by Pichler and Käber on Forest Informed Neural Networks, arXiv:2508.01228v2, with a different arXiv identifier, different authors, and different subject matter. It contains no mention of the MCP method, vision-language models, caches, prototypes, test-time adaptation, or the 15 downstream tasks advertised in the abstract. Consequently, there is no verifiable support for the central claim of the paper.
  2. [Abstract] The abstract claims that 'comparative and ablation experiments across 15 downstream tasks demonstrate that the proposed method and framework achieve state-of-the-art generalization performance,' but no such experiments, tables, figures, or numerical results are present anywhere in the provided record. Without any experimental detail, the claim of state-of-the-art performance cannot be checked.
  3. [Abstract] The paper's design rationale rests on the assertion that 'this study identifies a positive correlation between cache-enhanced performance and intra-class compactness.' No methodology, dataset, or analysis supporting this correlation is provided. This is a load-bearing observation for the proposed method, and its absence prevents any assessment of the method's motivation.
minor comments (2)
  1. [Project page] The abstract refers to a project page URL that may contain additional material, but the URL is not a substitute for the manuscript content and cannot be used to verify claims that should appear in the paper itself.
  2. [Metadata] The supplied arXiv identifier in the full text (2508.01228v2) differs from the advertised identifier (2508.01225), which suggests a mismatch that should be resolved before any editorial processing.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity can be established from the provided record; the supplied full text is an unrelated forest-ecology manuscript.

full rationale

The abstract describes a vision-language test-time adaptation method (MCP) with three caches and claims state-of-the-art performance over 15 downstream tasks. However, the supplied full text is a different paper, 'Inferring Processes Within Dynamic Forest Models Using Hybrid Modeling' by Pichler and Kaeber, with arXiv identifier 2508.01228v2 [q-bio.QM], and it does not mention caches, prototypes, test-time adaptation, or vision-language models. Because the actual method, equations, experiments, and ablations of the MCP paper are absent, there is no derivation chain to audit for circularity. From the abstract alone, no self-definitional step, no fitted-input-called-prediction, and no load-bearing self-citation chain can be exhibited. The claim that the method performs well against prior cache-enhanced methods is an empirical assertion, not an internal reduction. The mismatch is a serious completeness/correctness issue for review, but it is not evidence of circular reasoning. Under the hard rule requiring a specific quoted reduction or fitted-parameter-renamed-as-prediction, the honest finding is that no circularity is demonstrated.

Assumptions & free parameters 0 free parameters · 0 assumptions · 0 invented entities

No parameters, axioms, or invented entities are extracted because the supplied full text is a different manuscript and the MCP method is not described. The forest modeling paper contains model parameters, but it is not the paper under review.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Multi-Cache Enhanced Prototype Learning for Test-Time Generalization of Vision-Language Models." pith.science (2026). https://pith.science/paper/2DE7UKB6

@misc{pith2026250801225,
  author       = {Pith},
  title        = {Pith review of: Multi-Cache Enhanced Prototype Learning for Test-Time Generalization of Vision-Language Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2DE7UKB6}},
  note         = {Machine review of arXiv:2508.01225}
}
read the original abstract

In zero-shot setting, test-time adaptation adjusts pre-trained models using unlabeled data from the test phase to enhance performance on unknown test distributions. Existing cache-enhanced TTA methods rely on a low-entropy criterion to select samples for prototype construction, assuming intra-class compactness. However, low-entropy samples may be unreliable under distribution shifts, and the resulting prototypes may not ensure compact intra-class distributions. This study identifies a positive correlation between cache-enhanced performance and intra-class compactness. Based on this observation, we propose a Multi-Cache enhanced Prototype-based Test-Time Adaptation (MCP) featuring three caches: an entropy cache for initializing prototype representations with low-entropy samples, an align cache for integrating visual and textual information to achieve compact intra-class distributions, and a negative cache for prediction calibration using high-entropy samples. We further developed MCP++, a framework incorporating cross-modal prototype alignment and residual learning, introducing prototype residual fine-tuning. Comparative and ablation experiments across 15 downstream tasks demonstrate that the proposed method and framework achieve state-of-the-art generalization performance. Project Page available at: https://zhaihaotian.github.io/MCP-ICCV25/

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

81 extracted references · 54 canonical work pages

  1. [1]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION format.date year duplicate empty "emp...

  2. [2]

    Align your prompts: Test-time prompting with distribution alignment for zero-shot generalization

    Jameel Abdul Samadh, Mohammad Hanan Gani, Noor Hussein, Muhammad Uzair Khattak, Muhammad Muzammal Naseer, Fahad Shahbaz Khan, and Salman H Khan. Align your prompts: Test-time prompting with distribution alignment for zero-shot generalization. Advances in Neural Information Processing Systems, 36: 0 80396--80413, 2023

    2023

  3. [3]

    Dpa: Dual prototypes alignment for unsupervised adaptation of vision-language models

    Eman Ali, Sathira Silva, and Muhammad Haris Khan. Dpa: Dual prototypes alignment for unsupervised adaptation of vision-language models. In 2025 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), pages 6083--6093. IEEE, 2025

    work page 2025

  4. [4]

    Food-101--mining discriminative components with random forests

    Lukas Bossard, Matthieu Guillaumin, and Luc Van Gool. Food-101--mining discriminative components with random forests. In European Conference on Computer Vision, pages 446--461. Springer, 2014

    work page 2014

  5. [5]

    Proto-OOD: Enhancing OOD Object Detection with Prototype Feature Similarity

    Junkun Chen, Jilin Mei, Liang Chen, Fangzhou Zhao, Yan Xing, and Yu Hu. Proto-ood: Enhancing ood object detection with prototype feature similarity. arXiv preprint arXiv:2409.05466, 2024

    work page Pith review arXiv 2024

  6. [6]

    Describing textures in the wild

    Mircea Cimpoi, Subhransu Maji, Iasonas Kokkinos, Sammy Mohamed, and Andrea Vedaldi. Describing textures in the wild. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 3606--3613, 2013

    work page 2013

  7. [7]

    Imagenet: A large-scale hierarchical image database

    Jia Deng, Wei Dong, Richard Socher, Li-Jia Li, Kai Li, and Li Fei-Fei. Imagenet: A large-scale hierarchical image database. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 248--255. Ieee, 2009

    work page 2009

  8. [8]

    Learning to prompt for open-vocabulary object detection with vision-language model

    Yu Du, Fangyun Wei, Zihe Zhang, Miaojing Shi, Yue Gao, and Guoqi Li. Learning to prompt for open-vocabulary object detection with vision-language model. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 14064--14073, 2022

    work page 2022

Show all 81 references
  1. [9]

    Learning generative visual models from few training examples: An incremental bayesian approach tested on 101 object categories

    Li Fei-Fei, Rob Fergus, and Pietro Perona. Learning generative visual models from few training examples: An incremental bayesian approach tested on 101 object categories. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshop, pages 178--...

    2004

  2. [10]

    Diverse data augmentation with diffusions for effective test-time prompt tuning

    Chun-Mei Feng, Kai Yu, Yong Liu, Salman Khan, and Wangmeng Zuo. 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, 2023 a

    2023

  3. [11]

    Khan, and Wangmeng Zuo

    Chun-Mei Feng, Kai Yu, Yong Liu, Salman A. Khan, and Wangmeng Zuo. 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, 2023 b

    2023

  4. [12]

    Adapter merging with centroid prototype mapping for scalable class-incremental learning

    Takuma Fukuda, Hiroshi Kera, and Kazuhiko Kawamoto. Adapter merging with centroid prototype mapping for scalable class-incremental learning. In Proceedings of the Computer Vision and Pattern Recognition Conference, pages 4884--4893, 2025

    2025

  5. [13]

    Clip-adapter: Better vision-language models with feature adapters

    Peng Gao, Shijie Geng, Renrui Zhang, Teli Ma, Rongyao Fang, Yongfeng Zhang, Hongsheng Li, and Yu Qiao. Clip-adapter: Better vision-language models with feature adapters. International Journal of Computer Vision, 132 0 (2): 0 581--595, 2024

    2024

  6. [14]

    Queryable prototype multiple instance learning with vision-language models for incremental whole slide image classification

    Jiaxiang Gou, Luping Ji, Pei Liu, and Mao Ye. Queryable prototype multiple instance learning with vision-language models for incremental whole slide image classification. In Proceedings of the AAAI Conference on Artificial Intelligence, pages 3158--3166, 2025

    2025

  7. [15]

    Open-vocabulary object detection via vision and language knowledge distillation

    Xiuye Gu, Tsung-Yi Lin, Weicheng Kuo, and Yin Cui. Open-vocabulary object detection via vision and language knowledge distillation. In International Conference on Learning Representations, 2021

    2021

  8. [16]

    Huang, and Shi-Min Hu

    Meng-Hao Guo, Yi Zhang, Tai-Jiang Mu, Sharon X. Huang, and Shi-Min Hu. Tuning vision-language models with multiple prototypes clustering. IEEE Transactions on Pattern Analysis and Machine Intelligence, 46 0 (12): 0 11186--11199, 2024

    2024

  9. [17]

    Ranked entropy minimization for continual test-time adaptation

    Jisu Han, Jaemin Na, and Wonjun Hwang. Ranked entropy minimization for continual test-time adaptation. arXiv preprint arXiv:2505.16441, 2025

    2025 arXiv

  10. [18]

    Point-tta: Test-time adaptation for point cloud registration using multitask meta-auxiliary learning

    Ahmed Hatem, Yiming Qian, and Yang Wang. Point-tta: Test-time adaptation for point cloud registration using multitask meta-auxiliary learning. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 16494--16504, 2023

    2023

  11. [19]

    Eurosat: A novel dataset and deep learning benchmark for land use and land cover classification

    Patrick Helber, Benjamin Bischke, Andreas Dengel, and Damian Borth. 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 0 (7): 0 2217--2226, 2019

    2019

  12. [20]

    The many faces of robustness: A critical analysis of out-of-distribution generalization

    Dan Hendrycks, Steven Basart, Norman Mu, Saurav Kadavath, Frank Wang, Evan Dorundo, Rahul Desai, Tyler Zhu, Samyak Parajuli, Mike Guo, et al. The many faces of robustness: A critical analysis of out-of-distribution generalization. In In Proceedings of the IEEE/CVF Internationa...

    2021

  13. [21]

    The many faces of robustness: A critical analysis of out-of-distribution generalization

    Dan Hendrycks, Steven Basart, Norman Mu, Saurav Kadavath, Frank Wang, Evan Dorundo, Rahul Desai, Tyler Zhu, Samyak Parajuli, Mike Guo, et al. The many faces of robustness: A critical analysis of out-of-distribution generalization. In Proceedings of the IEEE/CVF international c...

    2021

  14. [22]

    Natural adversarial examples

    Dan Hendrycks, Kevin Zhao, Steven Basart, Jacob Steinhardt, and Dawn Song. Natural adversarial examples. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 15262--15271, 2021 c

    2021

  15. [23]

    Scaling up vision-language pretraining for image captioning

    Xiaowei Hu, Zhe Gan, Jianfeng Wang, Zhengyuan Yang, Zicheng Liu, Yumao Lu, and Lijuan Wang. Scaling up vision-language pretraining for image captioning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 17959--17968, 2021

    2021

  16. [24]

    Beyond entropy: Region confidence proxy for wild test-time adaptation

    Zixuan Hu, Yichun Hu, Xiaotong Li, Shixiang Tang, and Ling-Yu Duan. Beyond entropy: Region confidence proxy for wild test-time adaptation. arXiv preprint arXiv:2505.20704, 2025

    2025 arXiv

  17. [25]

    Le, Yun-Hsuan Sung, Zhen Li, and Tom Duerig

    Chao Jia, Yinfei Yang, Ye Xia, Yi-Ting Chen, Zarana Parekh, Hieu Pham, Quoc V. Le, Yun-Hsuan Sung, Zhen Li, and Tom Duerig. Scaling up visual and vision-language representation learning with noisy text supervision. In International Conference on Machine Learning, 2021

    2021

  18. [26]

    Few-shot object detection via feature reweighting

    Bingyi Kang, Zhuang Liu, Xin Wang, Fisher Yu, Jiashi Feng, and Trevor Darrell. Few-shot object detection via feature reweighting. 2019 IEEE/CVF International Conference on Computer Vision (ICCV), pages 8419--8428, 2018

    2019

  19. [27]

    Adilbek Karmanov, Dayan Guan, Shijian Lu, Abdulmotaleb El-Saddik, and Eric P. Xing. Efficient test-time adaptation of vision-language models. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pages 14162--14171, 2024

    2024

  20. [28]

    Khan, and Fahad Shahbaz Khan

    Muhammad Uzair Khattak, Hanoona Abdul Rasheed, Muhammad Maaz, Salman H. Khan, and Fahad Shahbaz Khan. Maple: Multi-modal prompt learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 19113--19122, 2022

    2022

  21. [29]

    3d object representations for fine-grained categorization

    Jonathan Krause, Michael Stark, Jia Deng, and Li Fei-Fei. 3d object representations for fine-grained categorization. In Proceedings of the IEEE International Conference on Computer Vision Workshops, pages 554--561, 2013

    2013

  22. [30]

    Entropy is not enough for test-time adaptation: From the perspective of disentangled factors

    Jonghyun Lee, Dahuin Jung, Saehyung Lee, Junsung Park, Juhyeon Shin, Uiwon Hwang, and Sungroh Yoon. Entropy is not enough for test-time adaptation: From the perspective of disentangled factors. arXiv preprint arXiv:2403.07366, 2024

    2024 arXiv

  23. [31]

    Dpu: Dynamic prototype updating for multimodal out-of-distribution detection

    Shawn Li, Huixian Gong, Hao Dong, Tiankai Yang, Zhengzhong Tu, and Yue Zhao. Dpu: Dynamic prototype updating for multimodal out-of-distribution detection. In Proceedings of the Computer Vision and Pattern Recognition Conference, pages 10193--10202, 2025

    2025

  24. [32]

    Promptkd: Unsupervised prompt distillation for vision-language models

    Zheng Li, Xiang Li, Xinyi Fu, Xin Zhang, Weiqiang Wang, Shuo Chen, and Jian Yang. Promptkd: Unsupervised prompt distillation for vision-language models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 26617--26626, 2024

    2024

  25. [33]

    A comprehensive survey on test-time adaptation under distribution shifts

    Jian Liang, Ran He, and Tieniu Tan. A comprehensive survey on test-time adaptation under distribution shifts. International Journal of Computer Vision, 133 0 (1): 0 31--64, 2025 a

    2025

  26. [34]

    Advancing reliable test-time adaptation of vision-language models under visual variations

    Yiwen Liang, Hui Chen, Yizhe Xiong, Zihan Zhou, Mengyao Lyu, Zijia Lin, Shuaicheng Niu, Sicheng Zhao, Jungong Han, and Guiguang Ding. Advancing reliable test-time adaptation of vision-language models under visual variations. arXiv preprint arXiv:2507.09500, 2025 b

    2025 arXiv

  27. [35]

    Probabilistic prototype calibration of vision-language models for generalized few-shot semantic segmentation

    Jie Liu, Jiayi Shen, Pan Zhou, Jan-Jakob Sonke, and Efstratios Gavves. Probabilistic prototype calibration of vision-language models for generalized few-shot semantic segmentation. arXiv preprint arXiv:2506.22979, 2025

    2025 arXiv

  28. [36]

    Learning with mixture of prototypes for out-of-distribution detection

    Haodong Lu, Dong Gong, Shuo Wang, Jason Xue, Lina Yao, and Kristen Moore. Learning with mixture of prototypes for out-of-distribution detection. arXiv preprint arXiv:2402.02653, 2024

    2024 arXiv

  29. [37]

    Fine-grained visual classification of aircraft

    Subhransu Maji, Esa Rahtu, Juho Kannala, Matthew Blaschko, and Andrea Vedaldi. Fine-grained visual classification of aircraft. arXiv preprint arXiv:1306.5151, 2013

    2013 arXiv

  30. [38]

    Black-box test-time prompt tuning for vision-language models

    Fan'an Meng, Chaoran Cui, Hongjun Dai, and Shuai Gong. Black-box test-time prompt tuning for vision-language models. In Proceedings of the AAAI Conference on Artificial Intelligence, pages 6099--6107, 2025

    2025

  31. [39]

    Multimodal prototype-enhanced network for few-shot action recognition

    Xinzhe Ni, Yong Liu, Hao Wen, Yatai Ji, Jing Xiao, and Yujiu Yang. Multimodal prototype-enhanced network for few-shot action recognition. arXiv preprint arXiv:2212.04873, 2022

    2022 arXiv

  32. [40]

    Automated flower classification over a large number of classes

    Maria-Elena Nilsback and Andrew Zisserman. Automated flower classification over a large number of classes. In Indian Conference on Computer Vision, Graphics and Image Processing, pages 722--729. IEEE, 2008

    2008

  33. [41]

    Proto-clip: Vision-language prototypical network for few-shot learning

    Kamalesh Palanisamy, Yu-Wei Chao, Xinya Du, Yu Xiang, et al. Proto-clip: Vision-language prototypical network for few-shot learning. In 2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 2594--2601. IEEE, 2024

    2024

  34. [42]

    Omkar M Parkhi, Andrea Vedaldi, Andrew Zisserman, and C. V. Jawahar. Cats and dogs. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 3498--3505. IEEE, 2012

    2012

  35. [43]

    What does a platypus look like? generating customized prompts for zero-shot image classification

    Sarah Pratt, Rosanne Liu, and Ali Farhadi. What does a platypus look like? generating customized prompts for zero-shot image classification. 2023 IEEE/CVF International Conference on Computer Vision (ICCV), pages 15645--15655, 2022

    2023

  36. [44]

    Bidirectional prototype-reward co-evolution for test-time adaptation of vision-language models

    Xiaozhen Qiao, Peng Huang, Jiakang Yuan, Xianda Guo, Bowen Ye, Zhe Sun, and Xuelong Li. Bidirectional prototype-reward co-evolution for test-time adaptation of vision-language models. arXiv preprint arXiv:2503.09394, 2025

    2025 arXiv

  37. [45]

    Learning clustering-based prototypes for compositional zero-shot learning

    Hongyu Qu, Jianan Wei, Xiangbo Shu, and Wenguan Wang. Learning clustering-based prototypes for compositional zero-shot learning. arXiv preprint arXiv:2502.06501, 2025

    2025 arXiv

  38. [46]

    Learning transferable visual models from natural language supervision

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, Gretchen Krueger, and Ilya Sutskever. Learning transferable visual models from natural language supervision. In International Con...

    2021

  39. [47]

    icarl: Incremental classifier and representation learning

    Sylvestre-Alvise Rebuffi, Alexander Kolesnikov, Georg Sperl, and Christoph H Lampert. icarl: Incremental classifier and representation learning. In Proceedings of the IEEE conference on Computer Vision and Pattern Recognition, pages 2001--2010, 2017

    2001

  40. [48]

    Do imagenet classifiers generalize to imagenet? In International Conference on Machine Learning, pages 5389--5400

    Benjamin Recht, Rebecca Roelofs, Ludwig Schmidt, and Vaishaal Shankar. Do imagenet classifiers generalize to imagenet? In International Conference on Machine Learning, pages 5389--5400. PMLR, 2019

    2019

  41. [49]

    One-shot learning for semantic segmentation

    Amirreza Shaban, Shray Bansal, Zhen Liu, Irfan Essa, and Byron Boots. One-shot learning for semantic segmentation. In Proceedings of the British Machine Vision Conference (BMVC), 2017

    2017

  42. [50]

    R-tpt: Improving adversarial robustness of vision-language models through test-time prompt tuning

    Lijun Sheng, Jian Liang, Zilei Wang, and Ran He. R-tpt: Improving adversarial robustness of vision-language models through test-time prompt tuning. arXiv preprint arXiv:2504.11195, 2025

    2025 arXiv

  43. [51]

    Test-time prompt tuning for zero-shot generalization in vision-language models

    Manli Shu, Weili Nie, De-An Huang, Zhiding Yu, Tom Goldstein, Anima Anandkumar, and Chaowei Xiao. Test-time prompt tuning for zero-shot generalization in vision-language models. arXiv preprint arXiv:2209.07511, 2022

    2022 arXiv

  44. [52]

    Prototypical networks for few-shot learning

    Jake Snell, Kevin Swersky, and Richard Zemel. Prototypical networks for few-shot learning. Advances in neural information processing systems, 30, 2017

    2017

  45. [53]

    Ucf101: A dataset of 101 human actions classes from videos in the wild

    K Soomro. Ucf101: A dataset of 101 human actions classes from videos in the wild. arXiv preprint arXiv:1212.0402, 2012

    2012 arXiv

  46. [54]

    Just shift it: Test-time prototype shifting for zero-shot generalization with vision-language models

    Elaine Sui, Xiaohan Wang, and Serena Yeung-Levy. Just shift it: Test-time prototype shifting for zero-shot generalization with vision-language models. arXiv preprint arXiv:2403.12952, 2024

    2024 arXiv

  47. [55]

    Point-cache: Test-time dynamic and hierarchical cache for robust and generalizable point cloud analysis

    Hongyu Sun, Qiuhong Ke, Ming Cheng, Yongcai Wang, Deying Li, Chenhui Gou, and Jianfei Cai. Point-cache: Test-time dynamic and hierarchical cache for robust and generalizable point cloud analysis. In Proceedings of the Computer Vision and Pattern Recognition, pages 1263--1275, 2025

    2025

  48. [56]

    Representation learning with contrastive predictive coding

    Aaron van den Oord, Yazhe Li, and Oriol Vinyals. Representation learning with contrastive predictive coding. arXiv preprint arXiv:1807.03748, 2018

    2018 arXiv

  49. [57]

    Attention is all you need

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, ukasz Kaiser, and Illia Polosukhin. Attention is all you need. Advances in Neural Information Processing Systems, 30, 2017

    2017

  50. [58]

    Olshausen, and Trevor Darrell

    Dequan Wang, Evan Shelhamer, Shaoteng Liu, Bruno A. Olshausen, and Trevor Darrell. Fully test-time adaptation by entropy minimization. arXiv preprint arXiv:2006.10726, 2020

    2006 arXiv

  51. [59]

    Ctpt: Continual test-time prompt tuning for vision-language models

    Fan Wang, Zhongyi Han, Xingbo Liu, Yilong Yin, and Xin Gao. Ctpt: Continual test-time prompt tuning for vision-language models. Pattern Recognition, 161: 0 111300, 2025

    2025

  52. [60]

    Learning robust global representations by penalizing local predictive power

    Haohan Wang, Songwei Ge, Zachary Lipton, and Eric P Xing. Learning robust global representations by penalizing local predictive power. Advances in neural information processing systems, 32, 2019 a

    2019

  53. [61]

    Learning robust global representations by penalizing local predictive power

    Haohan Wang, Songwei Ge, Zachary Lipton, and Eric P Xing. Learning robust global representations by penalizing local predictive power. Advances in Neural Information Processing Systems, 32, 2019 b

    2019

  54. [62]

    Deep visual domain adaptation: A survey

    Mei Wang and Weihong Deng. Deep visual domain adaptation: A survey. Neurocomputing, 312: 0 135--153, 2018

    2018

  55. [63]

    Sun database: Large-scale scene recognition from abbey to zoo

    Jianxiong Xiao, James Hays, Krista A Ehinger, Aude Oliva, and Antonio Torralba. Sun database: Large-scale scene recognition from abbey to zoo. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 3485--3492. IEEE, 2010

    2010

  56. [64]

    Tran, Yi Xu, Sampath Chanda, Liqun Chen, Belinda Zeng, Trishul M

    Jinyu Yang, Jiali Duan, S. Tran, Yi Xu, Sampath Chanda, Liqun Chen, Belinda Zeng, Trishul M. Chilimbi, and Junzhou Huang. Vision-language pre-training with triple contrastive learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages ...

    2022

  57. [65]

    Hasegawa-Johnson, Yingzhen Li, and Chang Dong Yoo

    Hee Suk Yoon, Eunseop Yoon, Joshua Tian Jin Tee, Mark A. Hasegawa-Johnson, Yingzhen Li, and Chang Dong Yoo. C-tpt: Calibrated test-time prompt tuning for vision-language models via text feature dispersion. arXiv preprint arXiv:2403.14119, 2024

    2024 arXiv

  58. [66]

    Task residual for tuning vision-language models

    Tao Yu, Zhihe Lu, Xin Jin, Zhibo Chen, and Xinchao Wang. Task residual for tuning vision-language models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 10899--10909, 2022

    2022

  59. [67]

    Mitigating cache noise in test-time adaptation for large vision-language models

    Haotian Zhai, Xinyu Chen, Can Zhang, Tianming Sha, and Ruirui Li. Mitigating cache noise in test-time adaptation for large vision-language models. arXiv preprint arXiv:2503.18334, 2025

    2025 arXiv

  60. [68]

    Sycara, and Yaqi Xie

    Ce Zhang, Simon Stepputtis, Katia P. Sycara, and Yaqi Xie. Dual prototype evolving for test-time generalization of vision-language models. arXiv preprint arXiv:2410.12790, 2024 a

    2024 arXiv

  61. [69]

    Sycara, and Yaqi Xie

    Ce Zhang, Simon Stepputtis, Katia P. Sycara, and Yaqi Xie. Enhancing vision-language few-shot adaptation with negative learning. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pages 5905--5915, 2024 b

    2024

  62. [70]

    Historical test-time prompt tuning for vision foundation models

    Jingyi Zhang, Jiaxing Huang, Xiaoqin Zhang, Ling Shao, and Shijian Lu. Historical test-time prompt tuning for vision foundation models. arXiv preprint arXiv:2410.20346, 2024 c

    2024 arXiv

  63. [71]

    Memo: Test time robustness via adaptation and augmentation

    Marvin Zhang, Sergey Levine, and Chelsea Finn. Memo: Test time robustness via adaptation and augmentation. Advances in Neural Information Processing Systems, 35: 0 38629--38642, 2022

    2022

  64. [72]

    Tip-adapter: Training-free clip-adapter for better vision-language modeling

    Renrui Zhang, Rongyao Fang, Wei Zhang, Peng Gao, Kunchang Li, Jifeng Dai, Yu Jiao Qiao, and Hongsheng Li. Tip-adapter: Training-free clip-adapter for better vision-language modeling. arXiv preprint arXiv:2111.03930, 2021

    2021 arXiv

  65. [73]

    Boostadapter: Improving vision-language test-time adaptation via regional bootstrapping

    Taolin Zhang, Jinpeng Wang, Hang Guo, Tao Dai, Bin Chen, and Shu-Tao Xia. Boostadapter: Improving vision-language test-time adaptation via regional bootstrapping. arXiv preprint arXiv:2410.15430, 2024 d

    2024 arXiv

  66. [74]

    Adanpc: Exploring non-parametric classifier for test-time adaptation

    Yifan Zhang, Xue Wang, Kexin Jin, Kun Yuan, Zhang Zhang, Liang Wang, Rong Jin, and Tieniu Tan. Adanpc: Exploring non-parametric classifier for test-time adaptation. In International Conference on Machine Learning, pages 41647--41676. PMLR, 2023 a

    2023

  67. [75]

    Unsupervised prototype adapter for vision-language models

    Yi Zhang, Ce Zhang, Xueting Hu, and Zhihai He. Unsupervised prototype adapter for vision-language models. In Chinese Conference on Pattern Recognition and Computer Vision (PRCV), pages 197--209. Springer, 2023 b

    2023

  68. [76]

    Node-adapter: Neural ordinary differential equations for better vision-language reasoning

    Yi Zhang, Chun-Wun Cheng, Ke Yu, Zhihai He, Carola-Bibiane Sch \"o nlieb, and Angelica I Aviles-Rivero. Node-adapter: Neural ordinary differential equations for better vision-language reasoning. arXiv preprint arXiv:2407.08672, 2024 e

    2024 arXiv

  69. [77]

    Concept-guided prompt learning for generalization in vision-language models

    Yi Zhang, Ce Zhang, Ke Yu, Yushun Tang, and Zhihai He. Concept-guided prompt learning for generalization in vision-language models. arXiv preprint arXiv:2401.07457, 2024 f

    2024 arXiv

  70. [78]

    Dual memory networks: A versatile adaptation approach for vision-language models

    Yabin Zhang, Wen-Qing Zhu, Hui Tang, Zhiyuan Ma, Kaiyang Zhou, and Lei Zhang. Dual memory networks: A versatile adaptation approach for vision-language models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 28718--28728, 2024 g

    2024

  71. [79]

    Learning to prompt for vision-language models

    Kaiyang Zhou, Jingkang Yang, Chen Change Loy, and Ziwei Liu. Learning to prompt for vision-language models. International Journal of Computer Vision, 130: 0 2337 -- 2348, 2021

    2021

  72. [80]

    Dynamic multimodal prototype learning in vision-language models

    Xingyu Zhu, Shuo Wang, Beier Zhu, Miaoge Li, Yunfan Li, Junfeng Fang, Zhicai Wang, Dongsheng Wang, and Hanwang Zhang. Dynamic multimodal prototype learning in vision-language models. arXiv preprint arXiv:2507.03657, 2025

    2025

  73. [81]

    Efficient test-time prompt tuning for vision-language models

    Yuhan Zhu, Guozhen Zhang, Chen Xu, Haocheng Shen, Xiaoxin Chen, Gangshan Wu, and Limin Wang. Efficient test-time prompt tuning for vision-language models. arXiv preprint arXiv:2408.05775, 2024

    2024 arXiv

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.