Pith. sign in

REVIEW 4 major objections 6 minor 1 cited by

Preserving AUC Fairness in Learning with Noisy Protected Groups

T0 review · 4 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read Noisy protected-group labels need not break AUC fairness

desk verdict The DRO-to-AUC extension is plausible and the experiments are consistent, but the TV bound γ is never shown to be valid, so the stated fairness guarantee doesn't bind. read the letter →

arxiv 2505.18532 v1 pith:N6ALJKWQ submitted 2025-05-24 cs.LG

Mingyang Wu , Li Lin , Wenbin Zhang , Xin Wang , Zhenhuan Yang , Shu Hu This is my paper
classification cs.LG
keywords AUCfairnessnoisyprotectedgroupsdistributionallyrobustoptimizationtotalvariationdistancegroupdeepfakedetectionpairwiseranking
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 tackles a gap in AUC-based fairness: existing methods assume protected attributes like gender and race are accurate, but real labels are often noisy, whether from survey response bias or from demographic attributes inferred on AI-generated faces. Its central claim is that a model trained with noisy protected-group labels can still guarantee fairness on the true, clean groups, up to a slack equal to the total variation distance between the clean and noisy group-conditional distributions. To achieve this, it casts AUC fairness as a distributionally robust optimization problem, enforces the fairness constraint over every distribution inside a TV-distance ball, and estimates the ball radius from data via a lemma relating the bound to the probability that clean and noisy group labels disagree. The paper presents the resulting algorithm and reports that it attains the lowest AUC fairness violation on tabular benchmarks with simulated label flips and on deepfake-detection benchmarks with naturally noisy demographic annotations.

What carries the argument

The load-bearing object is the pairwise AUC fairness function $g_{z,z'}(\theta)=AUC_{z,z'}(\theta)-AUC(\theta)$, which measures how far the ranking of positive examples from group $z$ against negative examples from group $z'$ deviates from the overall ranking. The method wraps this function in a distributionally robust constraint, requiring the violation to stay non-positive for every pairwise distribution $\tilde{p}_{z,z'}$ inside the $\ell^{1,1}$ ball $\{\tilde{p}:\|\tilde{p}-\hat{p}_{z,z'}\|_{1,1}\le 2\gamma_{z,z'}\}$. A minimax Lagrangian objective over the model parameters, the multipliers $\lambda_{z,z'}$, and the adversarial pairwise distributions turns the constraint into a trainable objective. The radius $\gamma_{z,z'}$ comes from Lemma 4.2, which bounds the total variation distance between clean and noisy group-conditional distributions by the conditional probability that the clean and noisy group labels disagree; in image experiments, this probability is estimated by prompting a vision-language model with `a photo of {group}` versus `a photo without {group}` and comparing cosine similarities.

What would settle it

On a deepfake test set with human-corrected demographic labels, count how often the vision-language model calls a pair clean when the protected label is actually flipped; if that disagreement exceeds the margin by which the true TV distance surpasses the selected $\gamma$, the Theorem 4.1 guarantee ceases to apply to the deployed model.

Watch

Extended reading notes

Core claim

The central claim is that training under noisy protected groups does not have to sacrifice AUC fairness. Theorem 4.1 shows: if a model satisfies the fairness constraints $\hat{g}_{z,z'}(\theta)\le 0$ for all noisy pairwise groups $(\hat{Z},\hat{Z}')=(z,z')$, then the clean-group fairness violation obeys $g_{z,z'}(\theta)\le\gamma_{z,z'}$, where $\gamma_{z,z'}$ is any upper bound on the total variation distance $TV(p_{z,z'},\hat{p}_{z,z'})$ between the clean and noisy group-conditional pairwise distributions. Lemma 4.2 turns that bound into a countable quantity, namely the probability that clean and noisy group labels disagree on a positive-negative pair, and the paper estimates it for images by prompting a vision-language model with group-specific phrases. On tabular benchmarks with simulated label flips and on deepfake benchmarks with naturally noisy demographic annotations, the method reports the lowest AUC fairness violation among compared methods, and it keeps the advantage on a human-corrected clean test set.

Load-bearing premise

The guarantee holds only if the estimated radius $\gamma$ really upper-bounds the total variation distance between clean and noisy group-conditional distributions; in the image experiments, $\gamma=0.02$ is computed from a vision-language model's agreement with group-specific prompts, with no proof that the model's verdicts bound the true mismatch probability and no human calibration.

Editorial extensions

If this is right

  • A model trained with noisy group labels can carry a certificate: for each pair of groups, the clean-group fairness violation is at most the estimated TV-ball radius $\gamma$.
  • Because the constraint is enforced for every distribution inside the TV ball, the method should keep the violation low even when the exact noise rate is unknown, as long as $\gamma$ covers the true shift; the paper's high-noise runs up to a 0.9 flip rate are consistent with this.
  • On deepfake detection, the method yields the lowest group-AUC gap among compared methods on four benchmarks, and it remains best on a human-corrected clean FF++ test set, indicating the improvement is not an artifact of noisy evaluation labels.
  • Removing the robustness constraint raises the violation, and removing the sharpness-aware minimization term raises it further, so both the DRO objective and the flatness-seeking optimization contribute to the reported fairness performance.

Reading between the lines

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

  • Because Lemma 4.2 assumes equal group priors between clean and noisy labels, a dataset whose noise mechanism changes group base rates, such as annotation errors concentrated in one demographic, would need a larger $\gamma$ than the estimated mismatch rate; that case is not tested in the paper.
  • A cheap calibration extension would be to compare the vision-language model's clean/noisy verdicts against a few hundred human-checked labels and inflate $\gamma$ by the observed disagreement, making the image guarantee robust to the estimator's own error.
  • The paper's tabular experiments choose $\gamma$ by grid search rather than by its own noise-estimation pipeline, and the paper states this as a limitation; a direct tabular analogue of the prompt-based estimator would let the same guarantee chain apply outside images.
  • Since the pairwise fairness function bounds every group pair against the overall AUC, the method implicitly controls both intra-group and inter-group disparities; an explicit two-metric bound could let practitioners trade them off separately.
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

4 major / 6 minor

Summary. The paper proposes a distributionally robust optimization (DRO) approach for AUC fairness when protected group labels are noisy. The authors define a group-level AUC fairness metric, formulate a constrained optimization problem with total variation (TV) distance constraints, and prove a theorem bounding the clean-group fairness violation by the TV distance between clean and noisy group-conditional distributions. They estimate the TV bound using a CLIP-based mismatch counter for images and use the known flip rate for tabular data. They design an SGDA algorithm with sharpness-aware minimization and evaluate on three tabular datasets and four deepfake detection datasets, reporting lower AUC fairness violation than several baselines. The central claim is that this is the first robust AUC fairness method with theoretical guarantees under noisy protected groups.

Significance. If the theoretical guarantee were instantiated correctly, the paper would be a useful contribution to the fairness literature, extending DRO-based robustness from classification losses to pairwise AUC metrics. The paper addresses a practical and underexplored problem, provides a clear formal adaptation of Wang et al. (2020), and includes extensive experiments covering multiple noise levels, high noise, multi-group settings, and a clean-label image benchmark. The code is made publicly available. However, the central guarantee is not actually connected to the experimental setup as described, and several assumption checks are missing. The empirical results are promising but the theoretical claims are currently overstated.

major comments (4)
  1. [Section 4.1, Eq. (5)] The definition of g_{z,z'}(θ) in Eq. (5) implicitly assumes that every pair consists of one positive and one negative sample; without this paired-sampling interpretation, E[h(θ)|Z=z,Z'=z'] does not equal AUC_{z,z'}(θ) - AUC(θ) because the conditional expectation includes the label indicators and the group proportions. The paper should state this sampling assumption explicitly, as the theoretical guarantee in Theorem 4.1 and the empirical objective in Eq. (9) both rely on it.
  2. [Section 5.1-5.2 and Lemma 4.2] The tabular experiments set the DRO radius γ to the per-sample flip rate, but Lemma 4.2 bounds TV by the conditional pairwise mismatch probability, which under independent flips is 1-(1-γ)^2 = 2γ-γ^2, strictly larger than γ. Moreover, Lemma 4.2 requires the marginal pairwise group probabilities to be unchanged by noise, which is not satisfied under uniform random flipping when the group prior is imbalanced (e.g., Bank has group ratio 0.05:1). As a result, Theorem 4.1 does not apply to the reported tabular results, and the central theoretical guarantee is not instantiated in the experiments.
  3. [Section 4.3, Eq. (10)] The CLIP-based estimate of γ in Eq. (10) is used as the TV upper bound for the image experiments, but the paper provides no evidence that CLIP's verdicts yield an upper bound on the true mismatch probability; indeed, the text concedes that CLIP predictions are not perfectly reliable. The sensitivity analysis in §5.3 selects γ by test-set violation, which is post-hoc and does not validate the bound. Without a calibration or a conservative correction, the guarantee in Theorem 4.1 cannot be claimed for the deployed image models.
  4. [Section 5.2 and Appendix F.2] The image benchmark results in Table 2 evaluate fairness using the noisy protected labels (as stated in Appendix F.2), whereas the paper's objective is clean-group fairness. The reported violations on DFDC, DFD, and Celeb-DF may therefore not reflect true clean-group disparity; only the FF++ clean-label evaluation in Table 7 addresses this, and it covers a single dataset. The cross-domain empirical claims in §5.2 are consequently not fully supported.
minor comments (6)
  1. [Abstract] The phrase 'optimizing AUC over accuracy' should be 'optimizing AUC instead of accuracy' to avoid ambiguity.
  2. [Section 5.2] For the Default dataset at noise level 0.1, 'reducing it by 5.14%' is ambiguous; the absolute reduction from 0.0701 to 0.0187 is 5.14 percentage points, which corresponds to about a 73% relative reduction. Please rephrase.
  3. [Appendix F.2] The sentence 'We added experiments with one more baseline to address your concern' is a reviewer-response artifact and should be removed from the manuscript.
  4. [Table 7] The Min/Max column header uses '↓' instead of '↑', which is inconsistent with the other tables.
  5. [Figure 3 and Section 5.3] The reported reductions of 3.40% and 1.70% do not match the gap values shown in the figure (e.g., 0.095 to 0.075 is about a 21% relative reduction); please reconcile the numbers.
  6. [Section 4.3] The sentence 'we regard the group label of the test image as clean' should read 'training image', since the estimation is performed on the training set.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the DRO guarantee is a conditional mathematical bound, and the TV-radius estimate is grounded in an external pre-trained model rather than in the paper's own conclusions.

full rationale

The paper's main claim, Theorem 4.1, is a direct consequence of the triangle inequality: for any model satisfying the noisy-group fairness constraint, the clean-group fairness gap is bounded by the total variation distance between the clean and noisy group-conditional distributions. This is a self-contained proof given in Appendix A.1, and it does not assume the conclusion. The DRO relaxation in Eq. (8) then enforces a stronger constraint over a TV ball, so feasibility implies the clean-group bound; this is standard minimax reasoning, not circular. The only externally estimated quantity is the TV radius gamma, obtained in Eq. (10) from CLIP-based mismatch counts. CLIP is a pre-trained model not trained on the target task, and the paper explicitly refrains from using CLIP for relabeling or classification, so the estimate is not fitted to the fairness metric being predicted. In tabular experiments, gamma is set equal to the known injected noise fraction, which is an experimental control rather than a fitted prediction; whether that choice actually satisfies the theorem's precondition for minority groups is a calibration concern, not a circularity. Self-citations to Lin et al. (2024, 2025), Ju et al. (2024), and Hu et al. (2023) appear for datasets, baselines, and inspiration, but they do not carry the load of the theoretical derivation, which is proved in the appendix. Accordingly, the derivation chain is not circular and no step reduces to its own inputs by construction.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

The method introduces no new physical or conceptual entities. The DRO ball is a mathematical construct. The main load-bearing assumptions are the validity of the CLIP-based noise estimate, the prior-marginal equality in Lemma 4.2, and the surrogate loss preservation.

free parameters (2)
  • γ (DRO radius per group pair) = Tabular: chosen from {0.1, 0.2, 0.3, 0.4, 0.5} per noise level; Image: fixed to 0.02 from CLIP estimate.
    The fairness guarantee in Theorem 4.1 holds only if γ upper-bounds the true TV distance. For tabular data the value is a hyperparameter tuned per noise level; for image data it is derived from a CLIP-based heuristic that is not proven to be an upper bound.
  • Sharpness perturbation radius ν = Tabular: from {0.0005, 0.001, 0.005}; Image: from {0.3, 0.5, 0.7}
    Controls the SAM perturbation magnitude; influences optimization and fairness-generalization but is not central to the fairness guarantee.
assumptions (5)
  • domain assumption Equal prior pairwise group probabilities for clean and noisy labels: P[(Z,Z')=(z,z')] = P[(\hat Z,\hat Z')=(z,z')]
    Used in Lemma 4.2 to equate the two conditional mismatch probabilities; not verified on real data.
  • domain assumption The chosen γ satisfies γ ≥ TV(p_{z,z'}, \hat p_{z,z'}) for all z,z'
    Theorem 4.1 requires this upper bound. The empirical estimate from Eq. (10) is a proxy and may under- or over-estimate.
  • ad hoc to paper CLIP cosine similarity between image and text prompts is a valid indicator of protected-group label correctness
    Used in Eq. (10) to count mismatches; no ground-truth validation is provided for the deepfake datasets.
  • domain assumption The surrogate logistic loss replaces the 0/1 indicator without changing the fairness optimum
    Standard in AUC optimization, but the fairness constraints are also replaced by a surrogate, which can change the feasible region.
  • domain assumption The paired sampling distribution treats all pairs as positive-negative, so E[h|Z,Z'] equals the AUC gap
    The population-level definition of the fairness metric relies on this interpretation.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Preserving AUC Fairness in Learning with Noisy Protected Groups." pith.science (2026). https://pith.science/paper/N6ALJKWQ

@misc{pith2026250518532,
  author       = {Pith},
  title        = {Pith review of: Preserving AUC Fairness in Learning with Noisy Protected Groups},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/N6ALJKWQ}},
  note         = {Machine review of arXiv:2505.18532}
}
read the original abstract

The Area Under the ROC Curve (AUC) is a key metric for classification, especially under class imbalance, with growing research focus on optimizing AUC over accuracy in applications like medical image analysis and deepfake detection. This leads to fairness in AUC optimization becoming crucial as biases can impact protected groups. While various fairness mitigation techniques exist, fairness considerations in AUC optimization remain in their early stages, with most research focusing on improving AUC fairness under the assumption of clean protected groups. However, these studies often overlook the impact of noisy protected groups, leading to fairness violations in practice. To address this, we propose the first robust AUC fairness approach under noisy protected groups with fairness theoretical guarantees using distributionally robust optimization. Extensive experiments on tabular and image datasets show that our method outperforms state-of-the-art approaches in preserving AUC fairness. The code is in https://github.com/Purdue-M2/AUC_Fairness_with_Noisy_Groups.

Figures

Figures reproduced from arXiv: 2505.18532 by the authors.

Figure 1
Figure 1. Illustrative inter-/intra-group AUC dicrepancy examples of existing MinimaxFairAUC method (Yang et al., 2023) (dashed curves) and our method (solid curves) on Default (Yeh & Lien, 2009) dataset with noisy levels 0 and 0.3, respectively. Notations are defined in Section 4. In general, our method is better than MinimaxFairAUC in preserving AUC fairness, demonstrating robustness to noisy groups. metric by comparing pos… view at source ↗
Figure 2
Figure 2. Impact of noisy protected group labels on AUC fairness violation (lower values indicate better AUC fairness) in two sce￾narios: (a) Socioeconomic Analysis and (b) Deepfake Detection. Mean value is shown in black line. The standard deviation is shown in blue background, where three random runs for each noise level. 3. Motivation To demonstrate the impact of noisy protected group levels on AUC fairness, we conduct exp… view at source ↗
Figure 3
Figure 3. (Left) Comparison of AUC gap on inter-group and intra-group across different datasets. For the tabular dataset, we compare our method with MinimaxFairAUC on the Bank dataset under a noise level of 0.1. For the image dataset, we compare ours with PG-FDD on Celeb-DF. (Right) AUC fairness violation across different γ values. γ has been set manually from {0.01, 0.02, 0.03, 0.04, 0.05}. Backbone Method DFDC DFD Robust SA… view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Efficiency frontier showing the trade-off between Average AUC and Average Fairness Violation across three tabular datasets (Adult, Bank, Default) at varying noise levels (0.1–0.3). 19 [PITH_FULL_IMAGE:figures/full_fig_p019_4.png]
Figure 5
Figure 5. Figure 5: Efficiency frontier showing the trade-off between Average AUC and Average Min/Max AUC across three tabular datasets (Adult, Bank, Default) at varying noise levels (0.1–0.3). 20 [PITH_FULL_IMAGE:figures/full_fig_p020_5.png]
Figure 6
Figure 6. Figure 6: Efficiency frontier showing the trade-off between detection performance (AUC) and fairness violation across four benchmark datasets (FF++, DFDC, DFD, Celeb-DF). 21 [PITH_FULL_IMAGE:figures/full_fig_p021_6.png]
Figure 7
Figure 7. Figure 7: Efficiency frontier showing the trade-off between detection performance (AUC) and Min/Max AUC across four benchmark datasets (FF++, DFDC, DFD, Celeb-DF). 22 [PITH_FULL_IMAGE:figures/full_fig_p022_7.png]

Discussion (0). Sign in to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Heterogeneous Ranking in Industrial-Scale Recommender Systems: A Case Study

    cs.IR 2026-07 conditional novelty 4.0 of 10

    Heterogeneity-conditioned gating and expert modulation improved multi-task ranking in Google Discover, narrowing the articles-vs-videos ranking gap and lifting feed engagement in A/B tests.

Reference graph

Works this paper leans on

59 extracted references · 46 canonical work pages · cited by 1 Pith paper

  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]

    https://www.kaggle.com/c/deepfake-detection-challenge

    Deepfake detection challenge. https://www.kaggle.com/c/deepfake-detection-challenge. Accessed: 2021-04-24

    work page 2021

  3. [3]

    Uci machine learning repository, 2007

    Asuncion, A., Newman, D., et al. Uci machine learning repository, 2007

    2007

  4. [4]

    H., et al

    Beutel, A., Chen, J., Doshi, T., Qian, H., Wei, L., Wu, Y., Heldt, L., Zhao, Z., Hong, L., Chi, E. H., et al. Fairness in recommendation ranking through pairwise comparisons. In Proceedings of the 25th ACM SIGKDD international conference on knowledge discovery & data mining, pp.\ 2212--2220, 2019

    work page 2019

  5. [5]

    and Haas, C

    Caton, S. and Haas, C. Fairness in machine learning: A survey. ACM Computing Surveys, 56 0 (7): 0 1--38, 2024

    2024

  6. [6]

    E., Huang, L., Keswani, V., and Vishnoi, N

    Celis, L. E., Huang, L., Keswani, V., and Vishnoi, N. K. Fair classification with noisy protected attributes: A framework with provable guarantees. In International Conference on Machine Learning, pp.\ 1349--1361. PMLR, 2021

    work page 2021

  7. [7]

    Xception: Deep learning with depthwise separable convolutions

    Chollet, F. Xception: Deep learning with depthwise separable convolutions. In Proceedings of the IEEE conference on computer vision and pattern recognition, pp.\ 1251--1258, 2017

    2017

  8. [8]

    and Mohri, M

    Cortes, C. and Mohri, M. Auc optimization vs. error rate minimization. Advances in neural information processing systems, 16, 2003

    work page 2003

Show all 59 references
  1. [9]

    Measuring and mitigating unintended bias in text classification

    Dixon, L., Li, J., Sorensen, J., Thain, N., and Vasserman, L. Measuring and mitigating unintended bias in text classification. In Proceedings of the 2018 AAAI/ACM Conference on AI, Ethics, and Society, pp.\ 67--73, 2018

    2018

  2. [10]

    S., and Pontil, M

    Donini, M., Oneto, L., Ben-David, S., Shawe-Taylor, J. S., and Pontil, M. Empirical risk minimization under fairness constraints. Advances in neural information processing systems, 31, 2018

    2018

  3. [11]

    Efficient projections onto the l 1-ball for learning in high dimensions

    Duchi, J., Shalev-Shwartz, S., Singer, Y., and Chandra, T. Efficient projections onto the l 1-ball for learning in high dimensions. In Proceedings of the 25th international conference on Machine learning, pp.\ 272--279, 2008

    2008

  4. [12]

    Duchi, J. C. and Namkoong, H. Learning models with uniform performance via distributionally robust optimization. The Annals of Statistics, 49 0 (3): 0 1378--1406, 2021

    2021

  5. [13]

    Edwards, D. A. On the kantorovich--rubinstein theorem. Expositiones Mathematicae, 29 0 (4): 0 387--398, 2011

    2011

  6. [14]

    Sharpness-aware minimization for efficiently improving generalization

    Foret, P., Kleiner, A., Mobahi, H., and Neyshabur, B. Sharpness-aware minimization for efficiently improving generalization. arXiv preprint arXiv:2010.01412, 2020

    2010 arXiv

  7. [15]

    C., and Schmidt, L

    Gardner, J., Perdomo, J. C., and Schmidt, L. Large scale transfer learning for tabular data via language modeling. arXiv preprint arXiv:2406.12031, 2024

    2024 arXiv

  8. [16]

    Measuring fairness of rankings under noisy sensitive information

    Ghazimatin, A., Kleindessner, M., Russell, C., Abedjan, Z., and Golebiowski, J. Measuring fairness of rankings under noisy sensitive information. In Proceedings of the 2022 ACM Conference on Fairness, Accountability, and Transparency, pp.\ 2263--2279, 2022

    2022

  9. [17]

    When fair classification meets noisy protected attributes

    Ghosh, A., Kvitca, P., and Wilson, C. When fair classification meets noisy protected attributes. In Proceedings of the 2023 AAAI/ACM Conference on AI, Ethics, and Society, pp.\ 679--690, 2023

    2023

  10. [18]

    Deepfakes dataset by google & jigsaw

    Google and Jigsaw. Deepfakes dataset by google & jigsaw. In https://ai.googleblog.com/2019/09/contributing-data-to-deepfakedetection.html , 2019

    2019

  11. [19]

    Robust attentive deep neural network for detecting gan-generated faces

    Guo, H., Hu, S., Wang, X., Chang, M.-C., and Lyu, S. Robust attentive deep neural network for detecting gan-generated faces. IEEE Access, 10: 0 32574--32583, 2022

    2022

  12. [20]

    M., and Wang, S

    Gupta, M., Cotter, A., Fard, M. M., and Wang, S. Proxy fairness. arXiv preprint arXiv:1806.11212, 2018. URL https://arxiv.org/abs/1806.11212

    2018 arXiv

  13. [21]

    Hanley, J. A. and McNeil, B. J. The meaning and use of the area under a receiver operating characteristic (roc) curve. Radiology, 143 0 (1): 0 29--36, 1982

    1982

  14. [22]

    Fairness without demographics in repeated loss minimization

    Hashimoto, T., Srivastava, M., Namkoong, H., and Liang, P. Fairness without demographics in repeated loss minimization. In International Conference on Machine Learning, pp.\ 1929--1938. PMLR, 2018

    1929

  15. [23]

    Dualcoop++: Fast and effective adaptation to multi-label recognition with limited annotations

    Hu, P., Sun, X., Sclaroff, S., and Saenko, K. Dualcoop++: Fast and effective adaptation to multi-label recognition with limited annotations. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2023

    2023

  16. [24]

    and Chen, G

    Hu, S. and Chen, G. H. Fairness in survival analysis with distributionally robust optimization. Journal of Machine Learning Research, 25 0 (246): 0 1--85, 2024

    2024

  17. [25]

    Mentornet: Learning data-driven curriculum for very deep neural networks on corrupted labels

    Jiang, L., Zhou, Z., Leung, T., Li, L.-J., and Fei-Fei, L. Mentornet: Learning data-driven curriculum for very deep neural networks on corrupted labels. In International conference on machine learning, pp.\ 2304--2313. PMLR, 2018

    2018

  18. [26]

    H., and Lyu, S

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

    2024

  19. [27]

    and Zhou, A

    Kallus, N. and Zhou, A. The fairness of risk scores beyond classification: Bipartite ranking and the xauc metric. Advances in neural information processing systems, 32, 2019

    2019

  20. [28]

    Assessing algorithmic fairness with unobserved protected class using data combination

    Kallus, N., Mao, X., and Zhou, A. Assessing algorithmic fairness with unobserved protected class using data combination. Management Science, 68 0 (3): 0 1959--1981, 2022

    1959

  21. [29]

    J., Kahou, S

    Kenfack, P. J., Kahou, S. E., and A \" vodji, U. A survey on fairness without demographics. Transactions on Machine Learning Research, 2024. ISSN 2835-8856. URL https://openreview.net/forum?id=3HE4vPNIfX

    2024

  22. [30]

    Domain adaptation explainability & fairness in ai for medical image analysis: Diagnosis of covid-19 based on 3-d chest ct-scans

    Kollias, D., Arsenos, A., and Kollias, S. Domain adaptation explainability & fairness in ai for medical image analysis: Diagnosis of covid-19 based on 3-d chest ct-scans. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.\ 4907--4914, 2024

    2024

  23. [31]

    Determinants of social desirability bias in sensitive surveys: a literature review

    Krumpal, I. Determinants of social desirability bias in sensitive surveys: a literature review. Quality & quantity, 47 0 (4): 0 2025--2047, 2013

    2025

  24. [32]

    Auc maximization under positive distribution shift

    Kumagai, A., Iwata, T., Takahashi, H., Nishiyama, T., and Fujiwara, Y. Auc maximization under positive distribution shift. In The Thirty-eighth Annual Conference on Neural Information Processing Systems

  25. [33]

    Fairness without demographics through adversarially reweighted learning

    Lahoti, P., Beutel, A., Chen, J., Lee, K., Prost, F., Thain, N., Wang, X., and Chi, E. Fairness without demographics through adversarially reweighted learning. Advances in neural information processing systems, 33: 0 728--740, 2020

    2020

  26. [34]

    C., and Sidford, A

    Levy, D., Carmon, Y., Duchi, J. C., and Sidford, A. Large-scale methods for distributionally robust optimization. Advances in Neural Information Processing Systems, 33: 0 8847--8860, 2020

    2020

  27. [35]

    Multimodal foundation models: From specialists to general-purpose assistants

    Li, C., Gan, Z., Yang, Z., Yang, J., Li, L., Wang, L., Gao, J., et al. Multimodal foundation models: From specialists to general-purpose assistants. Foundations and Trends in Computer Graphics and Vision , 16 0 (1-2): 0 1--214, 2024

    2024

  28. [36]

    Celeb-df: A new dataset for deepfake forensics

    Li, Y., Yang, X., Sun, P., Qi, H., and Lyu, S. Celeb-df: A new dataset for deepfake forensics. In CVPR, pp.\ 6,7, 2020

    2020

  29. [37]

    Preserving fairness generalization in deepfake detection

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

    2024

  30. [38]

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

    Lin, L., Santosh, Wu, M., Wang, X., and Hu, S. Ai-face: A million-scale demographically annotated ai-generated face dataset and fairness benchmark. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2025

    2025

  31. [39]

    and Vishnoi, N

    Mehrotra, A. and Vishnoi, N. Fair ranking with noisy protected attributes. Advances in Neural Information Processing Systems, 35: 0 31711--31725, 2022

    2022

  32. [40]

    Pairwise fairness for ranking and regression

    Narasimhan, H., Cotter, A., Gupta, M., and Wang, S. Pairwise fairness for ranking and regression. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 34, pp.\ 5248--5255, 2020

    2020

  33. [41]

    Learning a deep dual-level network for robust deepfake detection

    Pu, W., Hu, J., Wang, X., Li, Y., Hu, S., Zhu, B., Song, R., Song, Q., Wu, X., and Lyu, S. Learning a deep dual-level network for robust deepfake detection. Pattern Recognition, 130: 0 108832, 2022

    2022

  34. [42]

    W., Hallacy, C., Ramesh, A., Goh, G., Agarwal, S., Sastry, G., Askell, A., Mishkin, P., Clark, J., et al

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

    2021

  35. [43]

    Justice as fairness: A restatement

    Rawls, J. Justice as fairness: A restatement. Erin Kelly/Harvard University, 2001

    2001

  36. [44]

    T., Uryasev, S., et al

    Rockafellar, R. T., Uryasev, S., et al. Optimization of conditional value-at-risk. Journal of risk, 2: 0 21--42, 2000

    2000

  37. [45]

    Faceforensics++: Learning to detect manipulated facial images

    Rossler, A., Cozzolino, D., Verdoliva, L., Riess, C., Thies, J., and Nie ner, M. Faceforensics++: Learning to detect manipulated facial images. In Proceedings of the IEEE/CVF international conference on computer vision, pp.\ 1--11, 2019

    2019

  38. [46]

    and Le, Q

    Tan, M. and Le, Q. Efficientnet: Rethinking model scaling for convolutional neural networks. In International conference on machine learning, pp.\ 6105--6114. PMLR, 2019

    2019

  39. [47]

    M., Huang, H., Khan, M

    Tian, Y., Wen, C., Shi, M., Afzal, M. M., Huang, H., Khan, M. O., Luo, Y., Fang, Y., and Wang, M. Fairdomain: Achieving fairness in cross-domain medical image segmentation and classification. In European Conference on Computer Vision, pp.\ 251--271. Springer, 2025

    2025

  40. [48]

    Learning fair scoring functions: Bipartite ranking under roc-based fairness constraints

    Vogel, R., Bellet, A., and Cl \'e men c on, S. Learning fair scoring functions: Bipartite ranking under roc-based fairness constraints. In International conference on artificial intelligence and statistics, pp.\ 784--792. PMLR, 2021

    2021

  41. [49]

    Robust optimization for fairness with noisy protected groups

    Wang, S., Guo, W., Narasimhan, H., Cotter, A., Gupta, M., and Jordan, M. Robust optimization for fairness with noisy protected groups. Advances in neural information processing systems, 33: 0 5190--5203, 2020

    2020

  42. [50]

    Vision-language models are strong noisy label detectors

    Wei, T., Li, H.-T., Li, C., Shi, J.-X., Li, Y.-F., and Zhang, M.-L. Vision-language models are strong noisy label detectors. Advances in Neural Information Processing Systems, 37: 0 58154--58173, 2024

    2024

  43. [51]

    and Menon, A

    Williamson, R. and Menon, A. Fairness risk measures. In International conference on machine learning, pp.\ 6786--6797. PMLR, 2019

    2019

  44. [52]

    Deep auc maximization for medical image classification: Challenges and opportunities

    Yang, T. Deep auc maximization for medical image classification: Challenges and opportunities. arXiv preprint arXiv:2111.02400, 2021

    2021 arXiv

  45. [53]

    Algorithmic foundation of deep x-risk optimization

    Yang, T. Algorithmic foundation of deep x-risk optimization. arXiv preprint arXiv:2206.00439, 2022

    2022 arXiv

  46. [54]

    L., Varshney, K

    Yang, Z., Ko, Y. L., Varshney, K. R., and Ying, Y. Minimax auc fairness: Efficient algorithm with provable convergence. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 37, pp.\ 11909--11917, 2023

    2023

  47. [55]

    Stochastic methods for auc optimization subject to auc-based fairness constraints

    Yao, Y., Lin, Q., and Yang, T. Stochastic methods for auc optimization subject to auc-based fairness constraints. In International Conference on Artificial Intelligence and Statistics, pp.\ 10324--10342. PMLR, 2023

    2023

  48. [56]

    and Lien, C.-h

    Yeh, I.-C. and Lien, C.-h. The comparisons of data mining techniques for the predictive accuracy of probability of default of credit card clients. Expert systems with applications, 36 0 (2): 0 2473--2480, 2009

    2009

  49. [57]

    How does disagreement help generalization against label corruption? In International conference on machine learning, pp.\ 7164--7173

    Yu, X., Han, B., Yao, J., Niu, G., Tsang, I., and Sugiyama, M. How does disagreement help generalization against label corruption? In International conference on machine learning, pp.\ 7164--7173. PMLR, 2019

    2019

  50. [58]

    Large-scale robust deep auc maximization: A new surrogate loss and empirical studies on medical image classification

    Yuan, Z., Yan, Y., Sonka, M., and Yang, T. Large-scale robust deep auc maximization: A new surrogate loss and empirical studies on medical image classification. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pp.\ 3040--3049, 2021

    2021

  51. [59]

    Doubly robust auc optimization against noisy and adversarial samples

    Zhang, C., Shi, W., Luo, L., and Gu, B. Doubly robust auc optimization against noisy and adversarial samples. In Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, pp.\ 3195--3205, 2023

    2023

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.