Pith. sign in

REVIEW 4 major objections 5 minor 36 references

NEARL: Interacted Query Adaptation with Orthogonal Regularization for Medical Vision-Language Understanding

T0 review · 4 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read NEARL-CLIP claims that bidirectionally querying both modalities — image querying text and text querying image — plus orthogonal decoupling of adaptation increments, adapts CLIP to medical images with state-of-the-art accuracy on Pneumonia,

desk verdict Sensible adapter combination with consistent ablations, but the SOTA claim rests on cherry-picked binary subsets and missing training details. read the letter →

arxiv 2508.04101 v2 pith:VHNKIV7L submitted 2025-08-06 cs.CV

Zelin Peng , Yichen Zhao , Yu Huang , Piao Yang , Feilong Tang , Zhengqin Xu , Xiaokang Yang , Wei Shen This is my paper · ORCID
classification cs.CV
keywords medicalvision-languageunderstandingCLIPadaptationparameter-efficientfine-tuningcross-modalinteractionorthogonalregularizationpromptlearningimageclassificationmodalityalignment
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

NEARL-CLIP aims to show that CLIP's failure on medical images is mainly a modality-misalignment problem, not a capacity problem: adapting only the text branch or only one direction of interaction leaves the frozen image encoder unable to extract meaningful medical features. The paper proposes a bidirectional scheme in which learnable queries from each modality attend to the other's features at every layer (USEformer), and a second module (OCA) that decouples the new knowledge from the pretrained feature space via Gram-Schmidt orthogonalization to prevent interference. The reported results — 94.7% accuracy on chest X-ray Pneumonia, 92.6% on Alzheimer MRI, 98.5% on Retina OCT — beat prompt-learning and unidirectional baselines with only 1.46M trainable parameters, suggesting that mutual enrichment and feature decoupling together close the domain gap.

What carries the argument

Two coupled modules. USEformer (Unified Synergy Embedding Transformer): M stacked cross-attention blocks; learnable image queries attend to text features and text queries attend to image features (image-to-text and text-to-image attention), producing compact mutual enrichments that are fused back into both branches. OCA (Orthogonal Cross-Attention Adapter): for each layer, fuses the pretrained layer output with the USEformer output via cross-attention, then applies Gram-Schmidt orthogonalization, projecting the increment onto the orthogonal complement of the pretrained feature subspace: $\Delta f_\perp = \Delta f - \frac{\langle \Delta f, f\rangle}{\langle f, f\rangle} f$. This enforces geom

What would settle it

Run NEARL-CLIP and the best unidirectional baselines on the full multiclass versions of the same datasets — all four Alzheimer classes (including Very mild dementia) and all four OCT classes (CNV, DME, DRUSEN, NORMAL). If the accuracy gain over MaPLe or XCoOp shrinks or reverses once the discarded, harder classes are included, the central claim that bidirectional interaction resolves modality misalignment would be weakened.

Watch

Extended reading notes

Core claim

On its own terms, the paper establishes that dual-modality interaction with orthogonal regularization is a sufficient recipe for parameter-efficient CLIP adaptation to medical imaging. Each layer's frozen features are augmented by a compact cross-modal query pool: image queries summarize text features, text queries summarize image features, and the resulting increments are projected onto the orthogonal complement of the frozen feature subspace before being added. The paper reports best-in-table figures across three datasets of different modalities (X-ray, MRI, OCT), with accuracies of 94.7%, 92.6%, and 98.5% and F1-scores of 94.2%, 92.6%, and 98.2%, and attributes the gains to preventing the

Load-bearing premise

The evaluation rests on binary classification datasets created by discarding classes (e.g., Alzheimer's 'Very mild dementia' and Retina's CNV/DRUSEN classes), so the claim that bidirectional interaction closes the medical domain gap is only tested on the easiest split of each task.

Editorial extensions

If this is right

  • If the reported gains hold, CLIP-based medical classifiers can be built by freezing the backbone and training only ~1.5M parameters, which suits low-annotation medical settings.
  • The superiority over unidirectional baselines on all three datasets implies that one-way adaptation (e.g., image-to-text or text-to-image only) leaves systematic alignment errors that a symmetric querying mechanism can recover.
  • Orthogonal decoupling is presented as a generalisable component: any PEFT adapter can be post-processed with Gram-Schmidt projection to decorrelate new knowledge from pretrained features.
  • The framework's layer-wise analysis suggests mid-level layers carry the most useful cross-modal signal, implying future methods might concentrate interaction where it matters rather than across all layers.

Reading between the lines

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

  • Because the benchmarks are binary tasks created by discarding classes (Alzheimer drops 'Very mild dementia'; Retina keeps only DME vs NORMAL), the stated gains may not transfer to multi-class differential diagnosis, where the excluded classes are often the confusing ones.
  • A direct testable extension: apply the same orthogonal adapter to LoRA-style or prompt-only baselines to see whether the improvement comes specifically from bidirectional querying or from the decoupling regularizer alone.
  • The orthogonalization step is reminiscent of gradient-projection methods for continual learning; one could test whether NEARL-CLIP also reduces forgetting when adapting sequentially to multiple medical datasets.
  • Since the paper's layer-group experiment favours mid-level interaction, an adaptive-depth variant (learning where to insert USEformer blocks) might achieve similar accuracy with fewer parameters.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper proposes NEARL-CLIP, a parameter-efficient adaptation of CLIP for medical image classification. The method introduces two modules: USEformer, which performs bidirectional cross-modal interaction via learnable queries and cross-attention, and OCA, an adapter that applies Gram-Schmidt orthogonalization to project adaptation features onto the orthogonal complement of pre-trained features, claimed to separate 'truly novel' from 'incremental' knowledge. The model adds 1.46M learnable parameters and is evaluated on three medical datasets (Pneumonia, Alzheimer, Retina), reporting state-of-the-art accuracy and F1 against prompt-learning and unidirectional-interaction baselines. Ablations on Pneumonia show contributions of each module and sensitivity to hyperparameters M and r.

Significance. If the empirical claims hold, NEARL-CLIP offers a compact and effective recipe for adapting CLIP to medical domains, with a plausible mechanism (bidirectional interaction and orthogonal regularization) and strong parameter efficiency. The paper's strengths include a clean architecture, consistent internal ablations, and a clear attempt to preserve pre-trained generalization while enabling adaptation. However, the external validity of the headline SOTA claim is substantially weakened by the evaluation protocol: two of the three datasets are reduced to binary subsets by discarding diagnostically relevant classes, and no statistical significance testing is reported. The claimed conceptual distinction between 'truly novel' and 'incremental' knowledge is also not directly evidenced. These issues are fixable within the manuscript's scope, but they currently limit the strength of the conclusions.

major comments (4)
  1. [Sec. IV-A, Table I] The evaluation protocol reduces multi-class medical datasets to binary subsets: Alzheimer discards 'Very mild dementia' and randomly subsamples Non-demented to balance; Retina keeps only DME vs NORMAL, dropping CNV and DRUSEN. These are not arbitrary omissions: the discarded classes are diagnostic gradations where text-guided distinctions are most stressed. Table I therefore supports SOTA on curated binary subsets, not on the medical diagnosis tasks named in the abstract and conclusion. Please either evaluate on the full multi-class tasks, or provide a principled justification for excluding classes and add multi-class results to substantiate the 'medical vision-language understanding' claim.
  2. [Sec. IV-B, Table I] The headline improvement on Pneumonia is 2.1% over MaPLe (94.7±0.2 vs 92.6±3.2), but MaPLe's standard deviation is 3.2 and only three seeds are averaged. No significance tests are reported for any comparison or ablation. The claim of state-of-the-art performance is therefore not statistically established. Report pairwise significance tests (e.g., corrected t-test or bootstrap) or confidence intervals/effect sizes for the main comparisons and for the ablations in Tables II-IV.
  3. [Sec. IV-A, Implementation Details] The implementation details list architecture hyperparameters (M, Nq, Dq, r, tau, epoch) but omit the optimizer, learning rate, weight decay, batch size, input resolution, augmentations, and the exact prompt template. No code is released. Since all comparison methods were re-implemented (as stated in Table I), these details are necessary for reproducibility and for assessing whether the comparison is fair. Please provide the full training protocol.
  4. [Sec. III-C, Eq. (8)] Equation (8) is a fixed Gram-Schmidt projection of Δf onto the orthogonal complement of f. The paper's conceptual claim that this operation 'decouples new knowledge into truly novel and incremental components' is an interpretation, not something demonstrated by the experiments. The ablation shows that orthogonal regularization helps empirically, but it does not validate that the projected component corresponds to semantically 'truly novel' information. Please soften the interpretation or provide a direct analysis of the projected components (e.g., norms, similarity to pre-trained features, or downstream effect of each component) to support the claim.
minor comments (5)
  1. [Sec. IV-A, Retina] Typo: 'NROMAL' should be 'NORMAL'.
  2. [Abstract / Sec. V] The abstract says 'three medical imaging modalities (X-ray, MRI)' but the Retina dataset is OCT; the conclusion says '(X-ray, MRI, etc.)'. Please correct the modality list for consistency.
  3. [Sec. III-B, Eqs. (4)-(5)] The softmax expressions appear to have unmatched parentheses; clarify the numerator and denominator.
  4. [Sec. III-A] The notation 'The text encoder T processes prompts T' uses T for both the encoder and the prompt set, which is confusing. Consider different symbols.
  5. [Sec. IV-A] Seeds are averaged but no seed values or random split generation details are given; please specify how the three seeds were used.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: NEARL's components are defined by fixed equations and evaluated on held-out test splits; no result reduces to its inputs.

full rationale

The paper's central derivation is architectural: USEformer stacks cross-attention and FFN modules (Eqs. 4-5) and OCA applies a fixed Gram-Schmidt orthogonal projection (Eq. 8). None of these equations encodes or is fitted to the reported accuracies or F1 scores. The empirical claims are external benchmark results on fixed test splits, and the comparison methods are re-run under the same protocol; no parameter is fitted to the target metric. Self-citations, such as FATE [2], appear only as a comparison baseline and related work, not as a load-bearing premise for NEARL's design or for the SOTA claim. There is no uniqueness theorem imported from prior work, no ansatz justified solely by self-citation, and no renaming of a known result as an organizing principle. The dataset construction (discarding classes in Alzheimer and Retina) raises a validity concern, but it is not circularity because the reported results are measured on the resulting test sets rather than being derived from those choices by construction. The paper is self-contained against external benchmarks, so the circularity score is 0.

Assumptions & free parameters 2 free parameters · 4 assumptions · 1 invented entities

The central claim relies on a small number of tuned hyperparameters, a manual prompt template, and several domain assumptions about CLIP features and binary medical tasks. The invented entity is a conceptual decomposition rather than a physical one.

free parameters (2)
  • Architecture hyperparameters M, Nq, Dq, r, tau = M=6, Nq=32, Dq=128, r=8, tau=1e-2
    Chosen via ablations in Tables III and IV and validation; they affect capacity and final accuracy and are not derived from theory.
  • Prompt template = A [modality] of [CLASS]
    Manually selected medical prompt; baseline comparisons may use different prompt formats, making part of the gain format-dependent.
assumptions (4)
  • domain assumption CLIP's frozen ViT-B/16 and text encoder provide a sufficient base feature space for medical adaptation (Sec. IV-A, Implementation Details)
    The method only adds 1.46M parameters on top of frozen CLIP, so the whole approach assumes CLIP features carry enough transferable signal for medical tasks.
  • ad hoc to paper Medical classification can be reduced to the chosen binary per-dataset tasks (Sec. IV-A, Dataset)
    Alzheimer discards Very mild dementia, Retina discards two of four classes; this simplification is not justified as representative of real diagnostic workloads.
  • ad hoc to paper Gram-Schmidt projection of the adapter residual onto the orthogonal complement of the pretrained feature preserves generalization and captures 'truly novel' knowledge (Eq. 8, Sec. III-C)
    The knowledge-level interpretation of a geometric projection is asserted, not proven; only one ablation on Pneumonia supports it.
  • domain assumption Cross-attention with shared projection matrices across branches and layers provides sufficient bidirectionality (Eq. 4-5, Sec. III-B)
    The design assumes that symmetric image-to-text and text-to-image queries with tied weights are enough to align modalities; no comparison to untied or alternative interaction designs is given.
invented entities (1)
  • Orthogonally decoupled 'truly novel' vs 'incremental' knowledge components
    purpose: Conceptual justification for why the OCA projection improves adaptation
    The dichotomy is defined by a projection in feature space, not by an external measurement; only an ablation on one dataset supports the interpretation.

how reviews work

0 comments
Cite this review

Pith. "Pith review of NEARL: Interacted Query Adaptation with Orthogonal Regularization for Medical Vision-Language Understanding." pith.science (2026). https://pith.science/paper/VHNKIV7L

@misc{pith2026250804101,
  author       = {Pith},
  title        = {Pith review of: NEARL: Interacted Query Adaptation with Orthogonal Regularization for Medical Vision-Language Understanding},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VHNKIV7L}},
  note         = {Machine review of arXiv:2508.04101}
}
read the original abstract

Computer-aided medical image analysis is crucial for disease diagnosis and treatment planning. While vision-language models (VLMs) such as CLIP exhibit strong generalization ability, their direct application to medical imaging remains hindered by a substantial domain gap. Existing methods for bridging this gap, including prompt learning and unidirectional modality interaction, typically introduce domain knowledge into only one modality. However, such approaches fail to fully exploit CLIP's inherent dual-modality structure and overlook the synergistic effect of bidirectional cross-modal interaction, resulting in persistent modality misalignment. In this paper, we propose NEARL (iNteracted quEry Adaptation with oRthogonaL Regularization), a novel parameter-efficient VLM framework for bidirectional cross-modal interaction. NEARL consists of two key components: (1) the Unified Synergy Embedding Transformer (USEformer), which dynamically generates compact cross-modal queries to facilitate interaction; and (2) the Orthogonal Cross-Attention Adapter (OCA), which decouples new knowledge into truly novel and incremental components through orthogonal regularization. This design reduces interference from incremental components, enabling more focused learning of novel information and improving modality interaction in VLMs. Notably, NEARL introduces only 1.46M learnable parameters. Extensive experiments on three medical imaging modalities demonstrate state-of-the-art performance (e.g., a 2.3% relative improvement on the pneumonia dataset), along with fast inference and low memory overhead, highlighting its effectiveness for real-world medical vision-language understanding.

Figures

Figures reproduced from arXiv: 2508.04101 by the authors.

Figure 1
Figure 1. Comparative overview: previous methods vs. NEARL-CLIP. Previous [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Overview of the proposed NEARL-CLIP framework. It consists of two core innovations: (1) USEformer, which facilitates a fully bidirectional [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Unified Synergy Embedding Transformer (USEformer). In USEformer, [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Orthogonal Cross-Attention Adapter (OCA). OCA is a lightweight [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]
Figure 5
Figure 5. Figure 5: Layer-wise interaction analysis. Information interaction within isolated [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: Visualization of features generated by the pre-trained CLIP image encoder (a-c) and NEARL-CLIP image encoder (e-g) on the Pneumonia dataset. [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

36 extracted references · 25 canonical work pages

  1. [1]

    Learning to prompt for vision- language models,

    K. Zhou, J. Yang, C. C. Loy, and Z. Liu, “Learning to prompt for vision- language models,” International Journal of Computer Vision (IJCV) , 2022

    work page 2022

  2. [2]

    Fate: Feature-adapted parameter tuning for vision-language models,

    Z. Xu, Z. Peng, X. Yang, and W. Shen, “Fate: Feature-adapted parameter tuning for vision-language models,” in Proceedings of the AAAI Con- ference on Artificial Intelligence , vol. 39, no. 9, 2025, pp. 9014–9022

    work page 2025

  3. [3]

    Deco-net: Robust multimodal brain tumor segmentation via decoupled complementary knowledge distillation,

    Z. Zeng, Z. Peng, X. Yang, and W. Shen, “Deco-net: Robust multimodal brain tumor segmentation via decoupled complementary knowledge distillation,” in 2024 IEEE International Conference on Bioinformatics and Biomedicine (BIBM) . IEEE, 2024, pp. 2829–2836

    work page 2024

  4. [4]

    Pm 2: A new prompting multi-modal model paradigm for few-shot medical image classification,

    Z. Wang, Q. Sun, B. Zhang, W. Su, P. Wang, J. Zhang, and Q. Zhang, “Pm 2: A new prompting multi-modal model paradigm for few-shot medical image classification,” in 2024 IEEE International Conference on Bioinformatics and Biomedicine (BIBM) . IEEE, 2024, pp. 3799– 3804

    work page 2024

  5. [5]

    Swin-unet: Unet-like pure transformer for medical image segmenta- tion,

    H. Cao, Y . Wang, J. Chen, D. Jiang, X. Zhang, Q. Tian, and M. Wang, “Swin-unet: Unet-like pure transformer for medical image segmenta- tion,” in European conference on computer vision . Springer, 2022, pp. 205–218

    2022

  6. [6]

    Multi-modal masked autoencoders for medical vision-and-language pre-training,

    Z. Chen, Y . Du, J. Hu, Y . Liu, G. Li, X. Wan, and T.-H. Chang, “Multi-modal masked autoencoders for medical vision-and-language pre-training,” in International Conference on Medical Image Computing and Computer-Assisted Intervention . Springer, 2022, pp. 679–689

    work page 2022

  7. [7]

    Dm-gan: A data augmentation-based approach for imbalanced medical image classification,

    H. Ding, K. Zhang, and N. Huang, “Dm-gan: A data augmentation-based approach for imbalanced medical image classification,” in 2024 IEEE International Conference on Bioinformatics and Biomedicine (BIBM) . IEEE, 2024, pp. 3160–3165

    work page 2024

  8. [8]

    V oco: A simple-yet-effective volume contrastive learning framework for 3d medical image analysis,

    L. Wu, J. Zhuang, and H. Chen, “V oco: A simple-yet-effective volume contrastive learning framework for 3d medical image analysis,” in Pro- ceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2024, pp. 22 873–22 882

    work page 2024

Show all 36 references
  1. [9]

    Mim: Mask in mask self-supervised pre-training for 3d medical image analysis,

    J. Zhuang, L. Wu, Q. Wang, P. Fei, V . Vardhanabhuti, L. Luo, and H. Chen, “Mim: Mask in mask self-supervised pre-training for 3d medical image analysis,” IEEE Transactions on Medical Imaging, 2025

    2025

  2. [10]

    Learning transferable visual models from natural language supervision,

    A. Radford, J. W. Kim, C. Hallacy, A. Ramesh, G. Goh, S. Agarwal, G. Sastry, A. Askell, P. Mishkin, J. Clark et al., “Learning transferable visual models from natural language supervision,” in International conference on machine learning . PmLR, 2021, pp. 8748–8763

    2021

  3. [11]

    Xcoop: Explainable prompt learning for computer-aided diagnosis via concept-guided context op- timization,

    Y . Bie, L. Luo, Z. Chen, and H. Chen, “Xcoop: Explainable prompt learning for computer-aided diagnosis via concept-guided context op- timization,” in International Conference on Medical Image Computing and Computer-Assisted Intervention . Springer, 2024, pp. 773–783

    2024

  4. [12]

    Aligning medical images with general knowledge from large language models,

    X. Fang, Y . Lin, D. Zhang, K.-T. Cheng, and H. Chen, “Aligning medical images with general knowledge from large language models,” 2024

    2024

  5. [13]

    Gpt-4 technical report,

    J. Achiam, S. Adler, S. Agarwal, L. Ahmad, I. Akkaya, F. L. Aleman, D. Almeida, J. Altenschmidt, S. Altman, S. Anadkat et al. , “Gpt-4 technical report,” arXiv preprint arXiv:2303.08774 , 2023

    2023 arXiv

  6. [14]

    Conditional prompt learning for vision-language models,

    K. Zhou, J. Yang, C. C. Loy, and Z. Liu, “Conditional prompt learning for vision-language models,” in Proceedings of the IEEE/CVF confer- ence on computer vision and pattern recognition , 2022, pp. 16 816– 16 825

    2022

  7. [15]

    Maple: Multi-modal prompt learning,

    M. U. Khattak, H. Rasheed, M. Maaz, S. Khan, and F. S. Khan, “Maple: Multi-modal prompt learning,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2023, pp. 19 113–19 122

    2023

  8. [16]

    ¨Uber die aufl ¨osung linearer gleichungen mit unendlich vielen unbekannten,

    E. Schmidt, “ ¨Uber die aufl ¨osung linearer gleichungen mit unendlich vielen unbekannten,” Rendiconti del Circolo Matematico di Palermo (1884-1940), vol. 25, no. 1, pp. 53–77, 1908

    1940

  9. [17]

    Identifying medical diagnoses and treatable diseases by image-based deep learning,

    D. S. Kermany, M. Goldbaum, W. Cai, C. C. Valentim, H. Liang, S. L. Baxter, A. McKeown, G. Yang, X. Wu, F. Yan, and et al, “Identifying medical diagnoses and treatable diseases by image-based deep learning,” Cell, vol. 172, no. 5, pp. 1122–1131.e9, 2018

    2018

  10. [18]

    Flamingo: a visual language model for few-shot learning,

    J.-B. Alayrac, J. Donahue, P. Luc, A. Miech, I. Barr, Y . Hasson, K. Lenc, A. Mensch, K. Millican, M. Reynolds et al. , “Flamingo: a visual language model for few-shot learning,” Advances in neural information processing systems, vol. 35, pp. 23 716–23 736, 2022

    2022

  11. [19]

    Scaling up visual and vision-language representation learning with noisy text supervision,

    C. Jia, Y . Yang, Y . Xia, Y .-T. Chen, Z. Parekh, H. Pham, Q. Le, Y .-H. Sung, Z. Li, and T. Duerig, “Scaling up visual and vision-language representation learning with noisy text supervision,” in International conference on machine learning . PMLR, 2021, pp. 4904–4916

    2021

  12. [20]

    Blip: Bootstrapping language-image pre-training for unified vision-language understanding and generation,

    J. Li, D. Li, C. Xiong, and S. Hoi, “Blip: Bootstrapping language-image pre-training for unified vision-language understanding and generation,” in International conference on machine learning . PMLR, 2022, pp. 12 888–12 900

    2022

  13. [21]

    Language is not all you need: Aligning perception with language models,

    S. Huang, L. Dong, W. Wang, Y . Hao, S. Singhal, S. Ma, T. Lv, L. Cui, O. K. Mohammed, B. Patra et al. , “Language is not all you need: Aligning perception with language models,” Advances in Neural Information Processing Systems , vol. 36, pp. 72 096–72 109, 2023

    2023

  14. [22]

    Open-vocabulary detr with conditional matching,

    Y . Zang, W. Li, K. Zhou, C. Huang, and C. C. Loy, “Open-vocabulary detr with conditional matching,” in European conference on computer vision. Springer, 2022, pp. 106–122

    2022

  15. [23]

    Gridclip: One-stage object detection by grid-level clip representation learning,

    J. Lin, S. Sun, and S. Gong, “Gridclip: One-stage object detection by grid-level clip representation learning,” Pattern Recognition, p. 112187, 2025

    2025

  16. [24]

    Taskclip: Extend large vision-language model for task oriented object detection,

    H. Chen, W. Huang, Y . Ni, S. Yun, Y . Liu, F. Wen, A. Velasquez, H. Latapie, and M. Imani, “Taskclip: Extend large vision-language model for task oriented object detection,” in European Conference on Computer Vision. Springer, 2024, pp. 401–418

    2024

  17. [25]

    Weakclip: Adapting clip for weakly-supervised semantic segmentation,

    L. Zhu, X. Wang, J. Feng, T. Cheng, Y . Li, B. Jiang, D. Zhang, and J. Han, “Weakclip: Adapting clip for weakly-supervised semantic segmentation,” International Journal of Computer Vision, vol. 133, no. 3, pp. 1085–1105, 2025

    2025

  18. [26]

    Understanding fine-tuning clip for open-vocabulary semantic segmentation in hyperbolic space,

    Z. Peng, Z. Xu, Z. Zeng, C. Wen, Y . Huang, M. Yang, F. Tang, and W. Shen, “Understanding fine-tuning clip for open-vocabulary semantic segmentation in hyperbolic space,” in Proceedings of the Computer Vision and Pattern Recognition Conference , 2025, pp. 4562–4572

    2025

  19. [27]

    Parameter- efficient fine-tuning in hyperspherical space for open-vocabulary seman- tic segmentation,

    Z. Peng, Z. Xu, Z. Zeng, Y . Huang, Y . Wang, and W. Shen, “Parameter- efficient fine-tuning in hyperspherical space for open-vocabulary seman- tic segmentation,” in Proceedings of the Computer Vision and Pattern Recognition Conference, 2025, pp. 15 009–15 020

    2025

  20. [28]

    The power of scale for parameter-efficient prompt tuning,

    B. Lester, R. Al-Rfou, and N. Constant, “The power of scale for parameter-efficient prompt tuning,” arXiv preprint arXiv:2104.08691 , 2021

    2021 arXiv

  21. [29]

    Visual-language prompt tuning with knowledge-guided context optimization,

    H. Yao, R. Zhang, and C. Xu, “Visual-language prompt tuning with knowledge-guided context optimization,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2023, pp. 6757–6767

    2023

  22. [30]

    Advancing textual prompt learning with anchored attributes,

    Z. Li, Y . Song, M.-M. Cheng, X. Li, and J. Yang, “Advancing textual prompt learning with anchored attributes,” 2025. [Online]. Available: https://arxiv.org/abs/2412.09442

    2025 arXiv

  23. [31]

    Nlprompt: Noise-label prompt learning for vision-language models,

    B. Pan, Q. Li, X. Tang, W. Huang, Z. Fang, F. Liu, J. Wang, J. Yu, and Y . Shi, “Nlprompt: Noise-label prompt learning for vision-language models,” in Proceedings of the Computer Vision and Pattern Recognition Conference, 2025, pp. 19 963–19 973

    2025

  24. [32]

    Textrefiner: Internal visual feature as efficient refiner for vision-language models prompt tuning,

    J. Xie, Y . Zhang, J. Peng, Z. Huang, and L. Cao, “Textrefiner: Internal visual feature as efficient refiner for vision-language models prompt tuning,” 2024. [Online]. Available: https://arxiv.org/abs/2412.08176

    2024 arXiv

  25. [33]

    Open access series of imaging studies (oasis): Cross- sectional mri data in young, middle aged, nondemented, and demented older adults,

    D. S. Marcus, T. H. Wang, J. Parker, J. G. Csernansky, J. C. Morris, and R. L. Buckner, “Open access series of imaging studies (oasis): Cross- sectional mri data in young, middle aged, nondemented, and demented older adults,” Journal of Cognitive Neuroscience , vol. 19, no. 9,...

    2007

  26. [34]

    Lora: Low-rank adaptation of large language models

    E. J. Hu, Y . Shen, P. Wallis, Z. Allen-Zhu, Y . Li, S. Wang, L. Wang, W. Chen et al., “Lora: Low-rank adaptation of large language models.” ICLR, vol. 1, no. 2, p. 3, 2022

    2022

  27. [35]

    Principal component analysis,

    H. Abdi and L. J. Williams, “Principal component analysis,” Wiley interdisciplinary reviews: computational statistics, vol. 2, no. 4, pp. 433– 459, 2010

    2010

  28. [36]

    Visualizing data using t-sne,

    L. v. d. Maaten and G. Hinton, “Visualizing data using t-sne,” Journal of machine learning research , vol. 9, no. Nov, pp. 2579–2605, 2008

    2008

Pith tools

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