Pith. sign in

REVIEW 4 major objections 6 minor 56 references

Robust Fairness Vision-Language Learning for Medical Image Analysis

T0 review · 4 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read A training framework combines bad-pair mining with a Sinkhorn-distance fairness loss, improving equity-scaled AUC of medical vision-language models by up to 8.6%.

desk verdict A competent but overclaimed combination of two existing losses: the fairness claim is contradicted by the paper's own Table I, and the 8.6% headline is nowhere in the data. read the letter →

arxiv 2505.03153 v1 pith:42KZXRBD submitted 2025-05-06 cs.CV

Sparsh Bansal , Mingyang Wu , Xin Wang , Shu Hu This is my paper · ORCID
classification cs.CV
keywords fairnessvision-languagemodelsmedicalimageanalysiscontrastivelearningrobustnessSinkhorndistancebadpairminingglaucomadetection
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

This paper proposes a training framework that makes vision-language models for medical diagnosis both more accurate and fairer. The authors argue that two problems compound in medical VLMs: clinical notes summarized by large language models create noisy or faulty image-text pairs, and the models inherit demographic biases from pretraining. Their framework attacks both by modifying the contrastive loss during fine-tuning. A Dynamic Bad Pair Mining algorithm identifies and down-weights high- and low-loss batches, and a Sinkhorn-distance term penalizes divergence between the loss distribution of each protected group and the overall batch. On the FairVLMed glaucoma dataset, the resulting models outperform both the CLIP baseline and FairCLIP, with up to an 8.6% gain in equity-scaled AUC, a metric that rewards overall accuracy and penalizes gaps between demographic groups.

What carries the argument

The load-bearing mechanism is the interaction of two loss components. Dynamic Bad Pair Mining, adapted from time-series contrastive learning, estimates the mean $\mu_e$ and standard deviation $\sigma_e$ of historical per-batch losses and defines an interval $[a,b]$; batches outside it are down-weighted by a Gaussian reduction weight instead of removed. The Sinkhorn-distance fairness term $S_\epsilon(B_W,B_{W_\gamma})$ is an entropy-regularized optimal transport cost between the batch's loss distribution and each protected group's distribution, added with regularization coefficient $\lambda$. These two terms are combined into a final loss $L_3 = L_2 + \lambda \sum_\gamma S_\epsilon(B_W,B_{W_\gamma})$, where $L_2$ is the bad-pair-weighted contrastive loss.

What would settle it

Train the same model on a dataset whose batch losses are known to be strongly skewed or bimodal, for example by injecting a systematic proportion of corrupted image-text pairs, then check whether DBPM's down-weighting still improves equity-scaled AUC; if it does no better than randomly down-weighting the same fraction of batches, the Gaussian-threshold mechanism is not the cause. A simpler check is to compute a normality statistic on the recorded historical batch losses during training, since strong evidence against normality would undermine the threshold derivation.

Watch

Extended reading notes

Core claim

The central claim is that robustness and fairness can be added to any feature-based medical vision-language model through two loss-side modifications, without changing the backbone architecture. Training starts from the standard CLIP contrastive objective using symmetric cross-entropy. The framework records each batch's historical loss, assumes these per-batch losses are roughly Gaussian, and uses thresholds $a = \mu - \alpha \sigma$ and $b = \mu + \beta \sigma$ to mark batches as correct, noisy, or faulty; noisy and faulty batches receive reduced weights from a Gaussian kernel. It then adds a Sinkhorn-distance regularizer that measures, for each protected attribute such as race or gender, how far that group's similarity-score distribution sits from the overall batch distribution. The paper reports that this combined loss raises overall AUC (area under the ROC curve) from 63.68 to 70.84 for a CLIP ViT-B/16 backbone and improves equity-scaled AUC from 56.76 to 65.88, with similar gains when the same framework is applied to BLIP-2.

Load-bearing premise

The framework assumes the historical per-batch losses are Gaussian, so that the mean-and-standard-deviation thresholds $a$ and $b$ separate noisy batches from clean ones; if the loss distribution is skewed or bimodal, the down-weighting becomes arbitrary, and the method also operates on whole batches rather than individual image-text pairs.

Editorial extensions

If this is right

  • If the framework is correct, medical VLMs can be fine-tuned on large, noisy datasets without a fairness/accuracy trade-off: the reported numbers show accuracy and equity improving together.
  • Because the same loss modifications improved both CLIP and BLIP-2, the method is architecture-agnostic and can be bolted onto future feature-based VLMs.
  • The Sinkhorn regularizer visibly narrows group-wise AUC gaps for race and ethnicity, suggesting a route to more equitable glaucoma screening.
  • DBPM's batch-level down-weighting makes training more stable when clinical notes are LLM-summarized, removing the need for manual pair cleaning.
  • The framework's reported gains are measured by equity-scaled AUC, meaning improvements are counted only insofar as they hold across all protected groups.

Reading between the lines

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

  • Beyond the paper, the Gaussian assumption behind DBPM is directly testable: a goodness-of-fit test on the historical loss distribution during training would show whether the thresholds select noise or simply truncate a skewed distribution.
  • Beyond the paper, the fairness regularizer is currently applied per protected attribute separately; a joint or intersectional Sinkhorn term might close the remaining ethnicity and language gaps visible in the reported group-wise AUCs.
  • Beyond the paper, since DBPM weights whole batches rather than individual image-text pairs, a pair-level noisy-label model could be more precise; comparing the two on the same dataset would quantify how much granularity matters.
  • Beyond the paper, one could test the framework on a second disease or imaging modality; if the gains replicate, the combination of bad-pair down-weighting and distributional fairness is a general recipe for medical VLMs, not a glaucoma-specific fix.
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 / 6 minor

Summary. The paper proposes a framework for robust and fair vision-language learning in medical image analysis. It combines a Dynamic Bad Pair Mining (DBPM) algorithm, adapted from time-series contrastive learning, with a Sinkhorn-distance-based fairness regularizer inspired by FairCLIP, and applies this to fine-tune CLIP ViT-B/16 and BLIP-2 on the Harvard-FairVLMed glaucoma dataset. The authors report overall AUC improvements over the CLIP and FairCLIP baselines and claim up to an 8.6% improvement in equity-scaled AUC, concluding that the framework ensures robustness and fairness for VLMs in medical diagnosis.

Significance. The problem is timely and important: medical VLMs need to be robust to noisy text-image pairs and fair across demographic groups. The paper has useful ingredients: it uses a public benchmark, compares against two relevant baselines, tests two backbone architectures, and releases code. The overall AUC gains (67.6 to 70.84 for CLIP; 80.03 to 81.96 for BLIP-2) are nontrivial if reproducible. However, the central fairness claim is contradicted by the paper's own reported metrics, and the robustness contribution is not supported by ablations, uncertainty quantification, or a methodological justification of the batch-level Gaussian assumption. As it stands, the paper establishes at most an accuracy improvement from a particular combination of existing losses, not a robust and fair framework.

major comments (4)
  1. [Table I and Section V (Limitation)] Table I contradicts the central fairness claim. Relative to FairCLIP, Robust FairCLIP worsens Gender DPD from 0.35 to 2.71 and DEOdds from 5.54 to 7.08; Ethnicity DPD from 5.35 to 9.85, DEOdds from 8.88 to 14.00, and ES-AUC from 64.1 to 61.71; Language DEOdds from 16.3 to 23.77 and ES-AUC from 59.5 to 58.59. The only consistent gains are overall AUC and some group-wise AUCs, which are accuracy improvements, not fairness improvements. Moreover, the abstract's claim of "up to a 8.6% improvement when looking at equity-scaled AUC" is not supported by the ES-AUC column: the overall ES-AUC changes from 56.76 (CLIP) to 65.88 (Robust FairCLIP), a relative change of 16.1%, and from 65.7 (FairCLIP) to 65.88, a change of only 0.27%; no subgroup ES-AUC value in Table I yields 8.6%. Section V's own limitation statement admits that the fairness loss optimizes only one protected attribute at a time, which is consistent with the observed degradation on other attributes, yet the paper presents all four attributes as successes. This directly undermines the claim that the framework "ensures fairness."
  2. [Section IV-A (Experimental Settings) and Tables I-II] No uncertainty quantification is reported. All results are single-run point estimates without error bars or multiple seeds, and several protected groups are very small (for example, Spanish speakers are 1.7% and users of other languages 0.8% of the dataset). Differences in DPD and DEOdds of a few percentage points may be within noise, so the claimed improvements over FairCLIP and BLIP-2 cannot be assessed reliably. The authors should report means and standard deviations over at least 3-5 seeds and, where feasible, confidence intervals for the fairness metrics.
  3. [Section III-C (Dynamic Bad Pair Mining)] The DBPM component operates at the batch level, not at the individual image-text pair level, which contradicts the claim that it identifies "faulty image-text pairs." The historical loss S(i,e) is defined per batch, and the weights w(i,e) down-weight entire batches whose mean loss falls in the tails. A high-loss batch could contain many correct pairs and one very faulty pair, or vice versa; batch-level weighting cannot distinguish these cases. In addition, the Gaussian assumption for the distribution of batch losses is asserted without evidence or a goodness-of-fit check, so the thresholds in Eq. (2) are not justified. An ablation isolating the DBPM contribution and a sensitivity analysis over alpha and beta are needed.
  4. [Section III-D and Eq. (6)] The fairness term is not shown to control the reported fairness metrics. Eq. (6) adds a Sinkhorn distance between the distribution of positive-pair similarities for a protected group and the overall batch distribution, while the evaluation uses DPD and DEOdds computed from final predictions. No derivation or empirical analysis links this regularizer to the measured fairness metrics. Since the reported results show degraded fairness on several attributes, the choice of lambda = 1e-7 and the single-attribute form of the loss require justification and ablation studies.
minor comments (6)
  1. [Abstract and Section IV-B] The abstract attributes the 8.6% gain to equity-scaled AUC, while Section IV-B states the improvement is over total AUC; these statements are inconsistent and should be reconciled.
  2. [Section III-B] The notation is inconsistent: the dataset is defined as D = {x_i, y_i, d_i, p_i} with d_i as diagnosis and p_i as protected attributes, but later gamma_i is introduced for protected attributes, and Eq. (1) uses d_i as a scalar label inside the log term. Please standardize the notation.
  3. [Section III-D, Eq. (4)] Eq. (4) is typeset incorrectly: the summation over gamma and the distance function are garbled, and the objective as printed is not a valid optimization statement.
  4. [Abstract] The code link contains a space ("Robust Fairness for Medical Image.git") and is not a valid URL.
  5. [Throughout] There are numerous typos, including "Dymanic" in Section II-B, "a issue" in the Introduction, and "LLaV A" in Section II-A.
  6. [Table I caption] The caption states that bolded values represent the best value in the group, but Robust FairCLIP's worse DPD and DEOdds values are not bolded; please verify the bolding convention.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: the method is an empirical combination of external losses and external-baseline evaluations; the Table I fairness discrepancy is a correctness concern, not a circular derivation.

full rationale

No significant circularity. The proposed loss L3 = L2 + lambda * sum_gamma S_epsilon(BW, BW_gamma) is assembled from externally defined components: the symmetric cross-entropy contrastive loss L1 (Eq. 1), a reweighting term w(i,e) built from the historical mean and standard deviation of L1 (Eqs. 2-3), and the Sinkhorn-distance fairness term taken from FairCLIP [18] (Eqs. 5-6). None of these components is defined in terms of the paper's target claim, and no fitted parameter is renamed as a prediction; alpha, beta, lambda, and epsilon are fixed hyperparameters. The evaluations are against external baselines (CLIP, FairCLIP, BLIP-2) and the external Harvard-FairVLMed dataset, so the reported improvements are not forced by construction. The DBPM Gaussian-assumption concern and the mismatch between the '8.6% ES-AUC' wording and the table values are correctness and interpretation issues, not circularity; the paper's own Limitation paragraph also concedes that only one protected attribute is optimized at a time. The derivation chain is therefore self-contained rather than circular.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

The method's behavior is controlled by four hand-set hyperparameters. The bad pair mining relies on a Gaussian assumption on loss histories, and the fairness penalty assumes Sinkhorn distance is an appropriate measure of group disparity. No new entities are introduced.

free parameters (4)
  • alpha = 3
    Threshold on the number of standard deviations below the mean loss to flag noisy batches (low loss tail).
  • beta = 3
    Threshold above the mean loss to flag faulty batches (high loss tail).
  • epsilon = 1e-4
    Blur parameter for Sinkhorn divergence, controlling entropy regularization in the optimal transport distance.
  • lambda = 1e-7
    Regularization weight for the sum of Sinkhorn distances in the final loss.
assumptions (4)
  • ad hoc to paper The historical training loss distribution across batches is Gaussian.
    Section III-C states 'assuming it follows a Gaussian distribution' and uses mean and standard deviation to set thresholds. No empirical justification is provided.
  • domain assumption Sinkhorn distance between batch-level and group-level similarity distributions is a valid fairness regularizer.
    Inherited from FairCLIP [18]; the paper relies on this without re-deriving.
  • domain assumption Protected attribute labels in Harvard-FairVLMed are correct and complete.
    The analysis and fairness loss depend on these labels; the paper does not audit them.
  • domain assumption GPT-4 summarized clinical notes retain enough diagnostic information for glaucoma classification.
    The dataset uses these summaries as the text modality; the paper does not verify their quality.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Robust Fairness Vision-Language Learning for Medical Image Analysis." pith.science (2026). https://pith.science/paper/42KZXRBD

@misc{pith2026250503153,
  author       = {Pith},
  title        = {Pith review of: Robust Fairness Vision-Language Learning for Medical Image Analysis},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/42KZXRBD}},
  note         = {Machine review of arXiv:2505.03153}
}
read the original abstract

The advent of Vision-Language Models (VLMs) in medical image analysis has the potential to help process multimodal inputs and increase performance over traditional inference methods. However, when considering the domain in which these models will be implemented, fairness and robustness are important to ensure the model stays true for any patient. In this paper, we introduce a framework for ensuring robustness and fairness of VLM models. This framework modifies the loss function at training by identifying and adjusting faulty image-text pairs through a Dynamic Bad Pair Mining algorithm and also utilizing Sinkhorn distance to ensure the loss distributions of protected groups do not deviate from the total loss. Experimental testing of our framework shows up to a 8.6\% improvement when looking at equity-scaled AUC.

Figures

Figures reproduced from arXiv: 2505.03153 by the authors.

Figure 1
Figure 1. A few examples from the Harvard-FairVLMed database [18] that is used for training and fairness optimization, show￾ing SLO Fundus photography images for glaucoma diagnosis. Latino/Hispanic, Middle Eastern, and Indian respectively [17]. Therefore, accounting for racial biases in VLM models trained for the medical domain is a priority to ensure equity in the healthcare sector. However, the current domain focuses on gen… view at source ↗
Figure 2
Figure 2. Overview of our proposed model using CLIP ViT-B/16 [19] as the base model with the loss function algorithms to account for erroneous pairs through Bad Pair Mining and equity through the Sinkhorn distance tokens, but an empirical study shows that the effective length is closer to 20 tokens [39]. Therefore, modifications must be made, and generally these clinical notes are summarized through a general LLM such as GPT-… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

56 extracted references · 40 canonical work pages

  1. [1]

    A survey of medical vision-and-language applications and their techniques,

    Q. Chen, R. Zhao, S. Wang, V . M. H. Phan, A. v. d. Hengel, J. Verjans, Z. Liao, M.-S. To, Y . Xia, J. Chen, et al. , “A survey of medical vision-and-language applications and their techniques,” arXiv preprint arXiv:2411.12195, 2024

    arXiv 2024

  2. [2]

    Llm-medqa: Enhancing medical question answering through case studies in large language models,

    H. Yang, H. Chen, H. Guo, Y . Chen, C.-S. Lin, S. Hu, J. Hu, X. Wu, and X. Wang, “Llm-medqa: Enhancing medical question answering through case studies in large language models,” IJCNN, 2025

    2025

  3. [4]

    Robust clip-based detector for exposing diffusion model-generated images,

    L. Lin, I. Amerini, X. Wang, S. Hu, et al., “Robust clip-based detector for exposing diffusion model-generated images,” MIPR, 2024

    work page 2024

  4. [5]

    Uncertainty-aware explainable recommendation with large language models,

    Y . Peng, H. Chen, C.-S. Lin, G. Huang, J. Hu, H. Guo, B. Kong, S. Hu, X. Wu, and X. Wang, “Uncertainty-aware explainable recommendation with large language models,” in 2024 International Joint Conference on Neural Networks (IJCNN) , pp. 1–8, IEEE, 2024

    work page 2024

  5. [6]

    Detecting multimedia generated by large ai models: A survey,

    L. Lin, N. Gupta, Y . Zhang, H. Ren, C.-H. Liu, F. Ding, X. Wang, X. Li, L. Verdoliva, and S. Hu, “Detecting multimedia generated by large ai models: A survey,” arXiv preprint arXiv:2402.00045 , 2024

    arXiv 2024

  6. [7]

    Fg-smote: Towards fair node classification with graph neural network,

    Z. Wang, Z. Yin, Y . Zhang, L. Yang, T. Zhang, N. Pissinou, Y . Cai, S. Hu, Y . Li, L. Zhao,et al., “Fg-smote: Towards fair node classification with graph neural network,” ACM SIGKDD Explorations Newsletter , vol. 26, no. 2, pp. 99–108, 2025

    work page 2025

  7. [8]

    Towards fairness with limited demographics via disentan- gled learning,

    Z. Wang, A. Wu, N. Moniz, S. Hu, B. Knijnenburg, Q. Zhu, and W. Zhang, “Towards fairness with limited demographics via disentan- gled learning,” in Proceedings of the Thirty-fourth International Joint Conference on Artificial Intelligence , 2025

    work page 2025

  8. [9]

    Preserving auc fairness in learning with noisy protected groups,

    M. Wu, L. Lin, W. Zhang, X. Wang, Z. Yang, and S. Hu, “Preserving auc fairness in learning with noisy protected groups,” in The 42nd International Conference on Machine Learning (ICML) , 2025

    work page 2025

Show all 56 references
  1. [10]

    Ai-face: A million-scale demographically annotated ai-generated face dataset and fairness benchmark,

    L. Lin, X. Wang, S. Hu, et al., “Ai-face: A million-scale demographically annotated ai-generated face dataset and fairness benchmark,” CVPR, 2025

    2025

  2. [11]

    Preserving fairness generalization in deepfake detection,

    L. Lin, X. He, Y . Ju, X. Wang, F. Ding, and S. Hu, “Preserving fairness generalization in deepfake detection,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pp. 16815– 16825, 2024

    2024

  3. [12]

    Fairness in survival analysis with distributionally robust optimization,

    S. Hu and G. H. Chen, “Fairness in survival analysis with distributionally robust optimization,” Journal of Machine Learning Research , vol. 25, no. 246, pp. 1–85, 2024

    2024

  4. [13]

    Improving fairness in deepfake detection,

    Y . Ju, S. Hu, S. Jia, G. H. Chen, and S. Lyu, “Improving fairness in deepfake detection,” in Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision , pp. 4655–4665, 2024

    2024

  5. [14]

    Distributionally robust survival analysis: A novel fairness loss without demographics,

    S. Hu and G. H. Chen, “Distributionally robust survival analysis: A novel fairness loss without demographics,” in Machine Learning for Health , pp. 62–87, PMLR, 2022

    2022

  6. [15]

    Evaluation and mitigation of racial bias in clinical machine learning models: scoping review,

    J. Huang, G. Galal, M. Etemadi, and M. Vaidyanathan, “Evaluation and mitigation of racial bias in clinical machine learning models: scoping review,” JMIR Medical Informatics , vol. 10, no. 5, p. e36388, 2022

    2022

  7. [16]

    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 , pp. 8748–8763, PmLR, 2021

    2021

  8. [17]

    Identifying implicit social biases in vision-language models,

    K. Hamidieh, H. Zhang, W. Gerych, T. Hartvigsen, and M. Ghassemi, “Identifying implicit social biases in vision-language models,” in Pro- ceedings of the AAAI/ACM Conference on AI, Ethics, and Society, vol. 7, pp. 547–561, 2024

    2024

  9. [18]

    Fairclip: Harnessing fairness in vision-language learning,

    Y . Luo, M. Shi, M. O. Khan, M. M. Afzal, H. Huang, S. Yuan, Y . Tian, L. Song, A. Kouhana, T. Elze, et al. , “Fairclip: Harnessing fairness in vision-language learning,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pp. 12289–12301, 2024

    2024

  10. [19]

    An image is worth 16x16 words: Transformers for image recognition at scale,

    A. Dosovitskiy, L. Beyer, A. Kolesnikov, D. Weissenborn, X. Zhai, T. Unterthiner, M. Dehghani, M. Minderer, G. Heigold, S. Gelly, et al., “An image is worth 16x16 words: Transformers for image recognition at scale,” arXiv preprint arXiv:2010.11929 , 2020

    2010 arXiv

  11. [20]

    Robustly optimized deep feature decoupling network for fatty liver diseases detection,

    P. Huang, S. Hu, B. Peng, J. Zhang, X. Wu, and X. Wang, “Robustly optimized deep feature decoupling network for fatty liver diseases detection,” in International Conference on Medical Image Computing and Computer-Assisted Intervention , pp. 68–78, Springer, 2024

    2024

  12. [21]

    Diffusion-empowered autoprompt medsam,

    P. Huang, S. Hu, B. Peng, J. Zhang, H. Zhu, X. Wu, and X. Wang, “Diffusion-empowered autoprompt medsam,” arXiv preprint arXiv:2502.06817, 2025

    2025 arXiv

  13. [22]

    Uu- mamba: uncertainty-aware u-mamba for cardiac image segmentation,

    T. Y . Tsai, L. Lin, S. Hu, M.-C. Chang, H. Zhu, and X. Wang, “Uu- mamba: uncertainty-aware u-mamba for cardiac image segmentation,” in 2024 IEEE 7th International Conference on Multimedia Information Processing and Retrieval (MIPR) , pp. 267–273, IEEE, 2024

    2024

  14. [23]

    An explainable non-local network for covid-19 diagnosis,

    J. Yang, P. Huang, J. Hu, S. Hu, S. Lyu, X. Wang, J. Guo, and X. Wu, “An explainable non-local network for covid-19 diagnosis,” arXiv preprint arXiv:2408.04300 , 2024

    2024 arXiv

  15. [24]

    Robust covid-19 detection in ct images with clip,

    L. Lin, Y . S. Krubha, Z. Yang, C. Ren, T. D. Le, I. Amerini, X. Wang, and S. Hu, “Robust covid-19 detection in ct images with clip,” in 2024 IEEE 7th International Conference on Multimedia Information Processing and Retrieval (MIPR) , pp. 586–592, IEEE, 2024

    2024

  16. [25]

    Robust light-weight facial affective behavior recognition with clip,

    L. Lin, S. Papabathini, X. Wang, and S. Hu, “Robust light-weight facial affective behavior recognition with clip,” MIPR, 2024

    2024

  17. [26]

    U-medsam: Uncertainty-aware medsam for medical image segmentation,

    X. Wang, X. Liu, P. Huang, P. Huang, S. Hu, and H. Zhu, “U-medsam: Uncertainty-aware medsam for medical image segmentation,” arXiv preprint arXiv:2408.08881, 2024

    2024 arXiv

  18. [27]

    Repmedgraf: Re-parameterization medical generated radiation field for improved 3d image reconstruction,

    R. Sun, F. Liao, J. He, Y . Liu, W. Yi, Q. Fan, S. Hu, X. Wang, and J. Hu, “Repmedgraf: Re-parameterization medical generated radiation field for improved 3d image reconstruction,” in 2024 IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS) , p...

    2024

  19. [28]

    Cgd- net: A hybrid end-to-end network with gating decoding for liver tumor segmentation from ct images,

    X. Zhu, T. Liu, Z. Liu, O. Shaobo, X. Wang, S. Hu, and F. Ding, “Cgd- net: A hybrid end-to-end network with gating decoding for liver tumor segmentation from ct images,” in 2024 IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS) , pp. 1–7, IEEE, 2024

    2024

  20. [29]

    Umednerf: Uncertainty-aware single view volumetric rendering for medical neural radiance fields,

    J. Hu, Q. Fan, S. Hu, S. Lyu, X. Wu, and X. Wang, “Umednerf: Uncertainty-aware single view volumetric rendering for medical neural radiance fields,” in 2024 IEEE International Symposium on Biomedical Imaging (ISBI), pp. 1–4, IEEE, 2024

    2024

  21. [30]

    Neural radiance fields in medical imaging: Challenges and next steps,

    X. Wang, S. Hu, H. Fan, H. Zhu, and X. Li, “Neural radiance fields in medical imaging: Challenges and next steps,” arXiv preprint arXiv:2402.17797, 2024

    2024 arXiv

  22. [31]

    Attention guided policy optimization for 3d medical image registration,

    J. Hu, Z. Shuai, X. Wang, S. Hu, S. Sun, S. Lyu, and X. Wu, “Attention guided policy optimization for 3d medical image registration,” IEEE Access, vol. 11, pp. 65546–65558, 2023

    2023

  23. [32]

    Rank-based decomposable losses in machine learning: A survey,

    S. Hu, X. Wang, and S. Lyu, “Rank-based decomposable losses in machine learning: A survey,” IEEE Transactions on Pattern Analysis and Machine Intelligence , 2023

    2023

  24. [33]

    Pseudoprop: Robust pseudo-label generation for semi-supervised object detection in autonomous driving systems,

    S. Hu, C.-H. Liu, J. Dutta, M.-C. Chang, S. Lyu, and N. Ramakrish- nan, “Pseudoprop: Robust pseudo-label generation for semi-supervised object detection in autonomous driving systems,” in Proceedings of the 6 IEEE/CVF Conference on Computer Vision and Pattern Recognition , pp....

    2022

  25. [34]

    Robust attentive deep neural network for detecting gan-generated faces,

    H. Guo, S. Hu, X. Wang, M.-C. Chang, and S. Lyu, “Robust attentive deep neural network for detecting gan-generated faces,” IEEE Access , vol. 10, pp. 32574–32583, 2022

    2022

  26. [35]

    Hu, Rank-Based Losses in Machine Learning and Deep Learning

    S. Hu, Rank-Based Losses in Machine Learning and Deep Learning . PhD thesis, State University of New York at Buffalo, 2022

    2022

  27. [36]

    Sum of ranked range loss for supervised learning,

    S. Hu, Y . Ying, X. Wang, and S. Lyu, “Sum of ranked range loss for supervised learning,” Journal of Machine Learning Research , vol. 23, no. 112, pp. 1–44, 2022

    2022

  28. [37]

    Learning by minimizing the sum of ranked range,

    S. Hu, Y . Ying, X. Wang, and S. Lyu, “Learning by minimizing the sum of ranked range,” Advances in Neural Information Processing Systems , vol. 33, pp. 21013–21023, 2020

    2020

  29. [38]

    Robustness analysis of video-language models against visual and language pertur- bations,

    M. Schiappa, S. Vyas, H. Palangi, Y . Rawat, and V . Vineet, “Robustness analysis of video-language models against visual and language pertur- bations,” Advances in Neural Information Processing Systems , vol. 35, pp. 34405–34420, 2022

    2022

  30. [39]

    Long-clip: Unlocking the long-text capability of clip,

    B. Zhang, P. Zhang, X. Dong, Y . Zang, and J. Wang, “Long-clip: Unlocking the long-text capability of clip,” in European Conference on Computer Vision, pp. 310–325, Springer, 2024

    2024

  31. [40]

    Gpt-4 technical report. arxiv,

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

    2023 arXiv

  32. [41]

    Pmc-llama: toward building open-source language models for medicine,

    C. Wu, W. Lin, X. Zhang, Y . Zhang, W. Xie, and Y . Wang, “Pmc-llama: toward building open-source language models for medicine,” Journal of the American Medical Informatics Association, vol. 31, no. 9, pp. 1833– 1843, 2024

    2024

  33. [42]

    Prevalence of glaucoma in the united states: the 2005–2008 national health and nutrition examination survey,

    P. Gupta, D. Zhao, E. Guallar, F. Ko, M. V . Boland, and D. S. Friedman, “Prevalence of glaucoma in the united states: the 2005–2008 national health and nutrition examination survey,” Investigative ophthalmology & visual science, vol. 57, no. 6, pp. 2905–2913, 2016

    2005

  34. [43]

    Towards enhancing time series contrastive learning: A dynamic bad pair mining approach,

    X. Lan, H. Yan, S. Hong, and M. Feng, “Towards enhancing time series contrastive learning: A dynamic bad pair mining approach,” arXiv preprint arXiv:2302.03357, 2023

    2023 arXiv

  35. [44]

    Blip-2: Bootstrapping language- image pre-training with frozen image encoders and large language models,

    J. Li, D. Li, S. Savarese, and S. Hoi, “Blip-2: Bootstrapping language- image pre-training with frozen image encoders and large language models,” in International conference on machine learning , pp. 19730– 19742, PMLR, 2023

    2023

  36. [45]

    On large visual language models for medical imaging analysis: An empirical study,

    M.-H. Van, P. Verma, and X. Wu, “On large visual language models for medical imaging analysis: An empirical study,” in 2024 IEEE/ACM Conference on Connected Health: Applications, Systems and Engineer- ing Technologies (CHASE), pp. 172–176, IEEE, 2024

    2024

  37. [46]

    Transformers in medical image analysis,

    K. He, C. Gan, Z. Li, I. Rekik, Z. Yin, W. Ji, Y . Gao, Q. Wang, J. Zhang, and D. Shen, “Transformers in medical image analysis,” Intelligent Medicine, vol. 3, no. 1, pp. 59–78, 2023

    2023

  38. [47]

    Language augmentation in clip for improved anatomy detection on multi-modal medical images,

    M. Kakkar, D. Shanbhag, C. Aladahalli, and G. Reddy, “Language augmentation in clip for improved anatomy detection on multi-modal medical images,” in 2024 46th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) , pp. 1–4, IEEE, 2024

    2024

  39. [48]

    Eyeclip: A visual-language foundation model for multi-modal ophthalmic image analysis,

    D. Shi, W. Zhang, J. Yang, S. Huang, X. Chen, M. Yusufu, K. Jin, S. Lin, S. Liu, Q. Zhang, et al., “Eyeclip: A visual-language foundation model for multi-modal ophthalmic image analysis,” arXiv preprint arXiv:2409.06644, 2024

    2024 arXiv

  40. [49]

    Facmic: Federated adaptative clip model for medical image classification,

    Y . Wu, C. Desrosiers, and A. Chaddad, “Facmic: Federated adaptative clip model for medical image classification,” in International Conference on Medical Image Computing and Computer-Assisted Intervention , pp. 531–541, Springer, 2024

    2024

  41. [50]

    The reliability of llms for medical diagnosis: An exami- nation of consistency, manipulation, and contextual awareness,

    K. Subedi, “The reliability of llms for medical diagnosis: An exami- nation of consistency, manipulation, and contextual awareness,” arXiv preprint arXiv:2503.10647, 2025

    2025 arXiv

  42. [51]

    Contrastive learning with stronger augmenta- tions,

    X. Wang and G.-J. Qi, “Contrastive learning with stronger augmenta- tions,” IEEE transactions on pattern analysis and machine intelligence , vol. 45, no. 5, pp. 5549–5560, 2022

    2022

  43. [52]

    Rethinking positive pairs in contrastive learning,

    J. Wu, S. Mo, Z. Feng, S. Atito, J. Kitler, and M. Awais, “Rethinking positive pairs in contrastive learning,” arXiv preprint arXiv:2410.18200, 2024

    2024 arXiv

  44. [53]

    Vision language models in medicine,

    B. Chingnabe Kalpelbe, A. G. Adaambiik, and W. Peng, “Vision language models in medicine,” arXiv e-prints, pp. arXiv–2503, 2025

    2025

  45. [54]

    Fairness in medical image analysis and healthcare: A literature survey,

    Z. Xu, J. Li, Q. Yao, H. Li, and S. K. Zhou, “Fairness in medical image analysis and healthcare: A literature survey,” Authorea Preprints, 2023

    2023

  46. [55]

    Symmetric cross entropy for robust learning with noisy labels,

    Y . Wang, X. Ma, Z. Chen, Y . Luo, J. Yi, and J. Bailey, “Symmetric cross entropy for robust learning with noisy labels,” in Proceedings of the IEEE/CVF international conference on computer vision , pp. 322– 330, 2019

    2019

  47. [56]

    Computational optimal transport: With applications to data science,

    G. Peyr ´e, M. Cuturi, et al. , “Computational optimal transport: With applications to data science,” Foundations and Trends® in Machine Learning, vol. 11, no. 5-6, pp. 355–607, 2019

    2019

  48. [57]

    Harvard glaucoma fairness: a retinal nerve disease dataset for fairness learning and fair identity normalization,

    Y . Luo, Y . Tian, M. Shi, L. R. Pasquale, L. Q. Shen, N. Zebardast, T. Elze, and M. Wang, “Harvard glaucoma fairness: a retinal nerve disease dataset for fairness learning and fair identity normalization,” IEEE Transactions on Medical Imaging , 2024. 7

    2024

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.