On the Interplay of Human-AI Alignment,Fairness, and Performance Trade-offs in Medical Imaging

Aur\'elie Pahud de Mortanges; Haozhe Luo; Mauricio Reyes; Robert Berke; Zixin Shu; Ziyu Zhou

REVIEW 4 major objections 5 minor 1 cited by

On the Interplay of Human-AI Alignment,Fairness, and Performance Trade-offs in Medical Imaging

T0 review · 4 major / 5 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 The paper claims that steering a chest X-ray classifier's attention toward radiologist-annotated regions reduces fairness gaps across sex and age groups in most configurations, while also improving out-of-domain performance—provided the…

desk verdict First systematic study of human-AI alignment and fairness in chest X-ray, but the headline count doesn't reproduce and the random-attention control muddies the attribution; worth a careful revision, not a desk reject. read the letter →

arxiv 2505.10231 v1 pith:GYU2FAYC submitted 2025-05-15 cs.CV cs.AIcs.CL

Haozhe Luo , Ziyu Zhou , Zixin Shu , Aurélie Pahud de Mortanges , Robert Berke , Mauricio Reyes This is my paper · ORCID

classification cs.CVcs.AIcs.CL

keywords Human-AIalignmentfairnessmedicalimagingvision-languagemodelattentionchestX-rayout-of-domaingeneralizationexplanation-guidedlearning

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 asks whether teaching a chest X-ray classifier to look where radiologists look makes it fairer across demographic groups, not just more accurate. It reports that attention alignment reduces fairness gaps in 27 of 30 comparisons spanning three diseases and two demographic splits, and improves all four performance metrics on out-of-domain test sets. The benefit is strongest with partial alignment; full alignment can reverse gains. If true, this gives hospital AI developers a practical knob for fairness that does not require discarding accuracy.

What carries the argument

The machinery is a pre-trained medical vision-language model whose visual and text embeddings are fused by cross-attention into per-class attention maps; an Attention Aligner head supervises these maps with a modified dice loss (equation 1) that only penalizes positive pixels and suppresses false positives, added to the standard cross-entropy classification loss. This alignment loss is the only difference between baseline and aligned models, so fairness changes are attributed to it.

What would settle it

Train the same models with the alignment loss attached to random attention maps served at the same epochs, and compare fairness gains; if random maps produce the same fairness improvement as radiologist maps, then the content of the human guidance is not the active ingredient and the causal story fails. Alternatively, remove the attention loss and add an equivalent-magnitude regularizer such as dropout or weight decay; if fairness gains persist, the alignment mechanism itself is not responsible.

Watch

Extended reading notes

Core claim

The central discovery is that human-AI alignment—placing a loss on the model's cross-attention maps so they overlap with radiologist-annotated regions—acts as a fairness intervention in chest X-ray classification. Across edema, pleural effusion, and nodule/mass detection, aligned models show smaller accuracy, AUC, sensitivity, and F1 gaps between sex and age subgroups than the same model trained without alignment, while simultaneously improving out-of-domain performance. The paper also identifies a boundary condition: alignment is beneficial up to a point, and 100% alignment can degrade gains; randomized alignment destroys performance but still narrows fairness gaps, indicating that part of the fairness effect is a decorrelation effect rather than purely semantic guidance.

Load-bearing premise

The load-bearing premise is that the cross-attention maps the model produces honestly reflect the features driving its classification, so penalizing those maps to match radiologist annotations changes the learned representation rather than merely redrawing a visualization; if that link fails, the fairness gains could come from generic regularization rather than human guidance.

Editorial extensions

If this is right

Fairness gaps over sex and age shrink under attention alignment in the large majority of tested configurations (27 of 30).
Out-of-domain performance rises on all four metrics for all three diseases, so fairness gains do not have to come at the cost of generalization.
Full 100% alignment can erode the gains, so the dose of human guidance should be tuned rather than maximized.
In low-data settings, alignment reduces fairness gaps more strongly, making it a candidate intervention when training data is scarce.
Randomized alignment lowers performance while also lowering fairness gaps, which suggests that part of the fairness effect is a decorrelation effect independent of the semantic content of the human guidance.

Reading between the lines

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

Editorial inference: the same alignment mechanism might transfer beyond chest X-rays to other modalities with expert gaze maps, such as dermatology or pathology, where demographic skew also exists; the paper does not test this.
Editorial inference: because randomized alignment preserves fairness gains but kills performance, a practical extension would apply the alignment loss to a different layer or representation to isolate the generic regularizing component that drives the fairness benefit.
Editorial inference: a stricter multiple-comparison or permutation analysis across the 30 reported comparisons could reveal how many individual gains are statistically robust; the paper reports point estimates with error bars but no adjusted significance tests.

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.

Desk Editor's Note

read the letter

Two things to know up front. This is the first systematic study of whether explanation-guided / human-AI alignment training changes demographic fairness gaps in medical imaging, and that question is genuinely worth asking. The setup is substantial: three diseases, six public chest X-ray datasets, multiple alignment levels including a randomized-attention control, and both age and sex subgroups with a standard fairness-gap metric. If the effect holds, it's a useful training lever for groups working on fairness and robustness.

The problem is that the paper's own data are weaker than the abstract claims. The headline '27 out of 30 comparisons' doesn't reproduce from Table 2. Counting the four fairness-gap columns across the six dataset/subgroup rows gives 18 of 24 directional improvements; adding Hit Rate gives 24 of 30. So 'consistently reduces fairness gaps' is too strong, and the table-versus-abstract mismatch has to be fixed. No significance tests are reported anywhere, and with five repeats and overlapping error bars on several rows (JSRT nodule age accuracy, pleural effusion AUC), many of those 'improvements' could be noise.

The deeper issue is the randomized-attention control. The paper shows that random attention maps also reduce fairness gaps while degrading performance. That means the fairness benefit could be a generic decorrelation or regularization effect, not a consequence of encoding radiologist knowledge. The abstract says 'incorporating human insights consistently reduces fairness gaps,' but the authors' own control undermines the causal reading. A matched comparison with random guidance under equal loss weight and training budget would settle it.

None of this makes the paper worthless. The design is close to right, the authors ran the control that most papers skip, and the question is timely. But the claims need to be rebalanced, the counts need to be corrected, and statistical support is missing. I'd send this to peer review with a request for major revision, and I'd bring it to a reading group as a case study in how fairness interventions can be oversold. Not a desk reject, but nowhere near a clean accept.

Referee Report

4 major / 5 minor

Summary. The paper studies whether Human-AI alignment, implemented as supervision of a Vision Transformer's cross-attention maps with radiologist attention annotations (based on the authors' prior DWARF method), affects demographic fairness gaps in chest X-ray classification. It compares a no-alignment baseline (from KEVL, ref [30]) with an 'Aligner' model on three conditions and two out-of-domain datasets, across four fairness-gap metrics plus Hit Rate, with five repeats per experiment. The paper reports that alignment improves fairness in 27 of 30 comparisons, improves out-of-domain performance, helps in low-data regimes, and that random attention alignment also reduces fairness gaps while degrading performance. The central claim is that incorporating human insights 'consistently' reduces fairness gaps and improves generalization, with diminishing returns under excessive alignment.

Significance. The question is timely and the study is one of the first to connect explanation-guided learning to subgroup fairness in medical imaging. The empirical setup has clear strengths: multi-center training data (NIH, PadChest, VinDr, MIMIC), multiple OOD test sets, five repeats, a random-attention ablation, a low-data analysis, and public code. If the central claim were supported by a reproducible count and by statistical testing, the paper would be a solid contribution to both the fairness and explainability literatures. However, the headline '27/30' figure is not reproducible from Table 2, and the random-attention control suggests that the fairness benefit may not be specific to human knowledge. The significance is therefore conditional on a corrected and tightened analysis.

major comments (4)

[Abstract and Section 3, Table 2] The claim that Human-AI alignment improved fairness in '27 out of 30 comparisons' is not reproducible from Table 2 as printed. Counting only the four fairness-gap columns (Accuracy, AUC, Sensitivity, F1) across the six dataset-by-demographic-group rows gives 18 improvements and 6 worsening; if Hit Rate is counted as a fifth column, the total is 24 improvements, not 27. Hit Rate measures attention alignment rather than fairness, so it should not be pooled with fairness-gap improvements. This count is the main quantitative support for the 'consistently reduces fairness gaps' conclusion, so the number and the conclusion must be corrected.
[Section 2.1, Figure 4] The random-attention ablation is not a matched control for the attribution of the fairness effect to human insight. The text states only that random shapes at random locations are generated each epoch; it does not report the loss weight, the number or size of masks, or the training budget for the random condition, and Figure 4 presents results for the sex group only. Since random alignment also reduces fairness gaps, the current evidence cannot distinguish human semantic guidance from generic attention-map regularization or decorrelation. A matched comparison with equal loss weight and training budget, for both sex and age groups, is needed before the effect can be attributed to Human-AI alignment.
[Section 3, Tables and Figures] No statistical significance tests are reported, although each experiment was repeated five times. Several printed means and standard deviations overlap substantially, e.g., CheXpert Pleural Effusion Gender AUC gap is 0.21±0.18 for w/o and 0.23±0.20 for Aligner, and JSRT Nodule Age accuracy gap is 16.11±2.42 versus 20.09±4.07. Without paired hypothesis tests or confidence intervals, the statement that alignment 'consistently reduces fairness gaps' (Abstract and Section 4) is stronger than the evidence supports. Figure 5 also reports low-data comparisons without visible or described error bars, making it impossible to assess whether the displayed differences are meaningful.
[Section 3, Figure 3] Figure 3 and the surrounding text report that fairness improves up to 75% alignment but degrades at 100%, yet Table 2 labels the fully aligned model as 'Aligner' and summarizes it with a universal improvement count. The paper should state explicitly which alignment level is used in Table 2 and in Figures 5 and 6, and it should reconcile the non-monotonic trend in Figure 3 with the 'consistently reduces fairness gaps' phrasing used in the Abstract and Conclusion.

minor comments (5)

[Section 2.2, Equation (1)] Equation (1) introduces α, ε, and w_FP, but only w_FP = 2.0 is reported; the smoothing-term values for α and ε should be given.
[Introduction and Section 2.2] The Introduction cites [21] twice in the list '[5,6,7,21,26,28,21]', and Section 2.2 contains the typo 'derive the reader to [18]' instead of 'refer the reader to [18]'.
[Figure 5] Figure 5 reports comparisons across four training-data ratios from repeated runs, but no error bars are visible or described in the caption, so the stability of the displayed differences cannot be assessed.
[Figure 6] Figure 6 shows radar charts without numerical values or error bars, making the claim of 'considerable performance improvements' difficult to verify from the figure alone.
[Table 1] Table 1 prints 'JSR T' in the column heading, and CheXlocalize is listed as an additional attention evaluation set, but no attention-localization evaluation results are reported in the paper.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the fairness gaps and OOD performance are measured on external benchmark data, and the alignment loss is a training objective rather than a fitted predictor of the reported outcomes.

full rationale

This is an empirical benchmarking study, not a derivation with a claimed prediction. Equation (1) defines the attention-alignment training loss, and the final loss L_total = L_CE + L_AL is used only to train the model. The reported fairness-gap values are computed on out-of-distribution CheXpert and JSRT data using the external fairness-gap metric of Ktena et al. [15], with no parameter fitted to those fairness outcomes. The baseline 'w/o' model is from Zhang et al. [30], an external group, while the Aligner builds on the authors' prior VLM work [18]; that self-citation is a method-reuse disclosure rather than load-bearing circularity, since the central empirical claims are evaluated on data and metrics outside the training objective. The randomized-alignment ablation in Fig. 4 is a legitimate control and may weaken the causal attribution to human semantics, but that is an evidence-strength issue, not a definitional circularity. I additionally note that the paper's headline count '27 out of 30 comparisons' is not reproducible from the printed Table 2: counting the four fairness-gap columns across six rows gives 18 of 24 directional improvements, and adding the six Hit Rate improvements gives 24 of 30, not 27. This is a reporting inconsistency or possible counting error, not a circular reduction, so it does not raise the circularity score. Overall, the derivation chain is self-contained against external benchmarks and the fairness results are not forced by construction.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The central claim rests on the validity of the attention alignment mechanism, the quality of the radiologist annotations, and the chosen fairness metric. These are domain assumptions, not derived results.

free parameters (2)

False-positive suppression weight w_FP = 2.0
Chosen by hand following [18], used in Eq. 1; it controls the penalty on false-positive attention and could influence the fairness results.
Dice loss smoothing terms alpha and epsilon = not reported
Small constants in Eq. 1; values are not stated in the paper, so the exact loss is not fully reproducible from the text.

assumptions (4)

domain assumption Pretrained ViT-B and Med-KEBERT encoders provide suitable representations for the classification and alignment tasks.
Section 2.2 relies on these pretrained models without validating that their attention maps are meaningful for supervision.
domain assumption Human attention annotations from VinDr-CXR and CheXlocalize are accurate and appropriate as training targets.
Table 1 and Eq. 1 treat these annotations as ground truth for attention; their quality is not assessed in the paper.
domain assumption The fairness gap metric, defined as the AUC difference between best- and worst-performing subgroups, is a valid operationalization of fairness.
Section 2.1 follows [15] without discussing alternative fairness definitions that might change the conclusions.
ad hoc to paper Cross-attention maps from the VLM faithfully represent the features driving the classifier.
Section 2.2 assumes that supervising cross-attention to match human maps shapes the learned representation; this core mechanism is not independently verified.

how reviews work

0 comments

Cite this review

Pith. "Pith review of On the Interplay of Human-AI Alignment,Fairness, and Performance Trade-offs in Medical Imaging." pith.science (2026). https://pith.science/paper/GYU2FAYC

@misc{pith2026250510231,
  author       = {Pith},
  title        = {Pith review of: On the Interplay of Human-AI Alignment,Fairness, and Performance Trade-offs in Medical Imaging},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/GYU2FAYC}},
  note         = {Machine review of arXiv:2505.10231}
}

read the original abstract

Deep neural networks excel in medical imaging but remain prone to biases, leading to fairness gaps across demographic groups. We provide the first systematic exploration of Human-AI alignment and fairness in this domain. Our results show that incorporating human insights consistently reduces fairness gaps and enhances out-of-domain generalization, though excessive alignment can introduce performance trade-offs, emphasizing the need for calibrated strategies. These findings highlight Human-AI alignment as a promising approach for developing fair, robust, and generalizable medical AI systems, striking a balance between expert guidance and automated efficiency. Our code is available at https://github.com/Roypic/Aligner.

Figures

Figures reproduced from arXiv: 2505.10231 by the authors.

**Figure 2.** Human-AI Alignment flow chart adapted from [18]. The approach is based [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

**Figure 3.** Fairness-performance trade-off for age and sex groups across five levels [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗

Figures from the paper (2 more)

**Figure 4.** Figure 4: Fairness-performance trade-off results under randomized Human-AI alignment (green) for sex group demographics, compared to a baseline model without alignment (blue), and with Human-AI alignment (red). Error bars represent variability. 25% 50% 75% 100% Training Data R…

**Figure 6.** Figure 6: Effect of Human-AI alignment on out-of-domain samples for nodule de [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]

Discussion (0). Continue with ORCID 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

LVPNet: A Latent-variable-based Prediction-driven End-to-end Framework for Lossless Compression of Medical Images
eess.IV 2025-06 conditional novelty 5.0 of 10

LVPNet reports lower bits-per-pixel than prior learned lossless codecs by conditioning pixel predictions on a global multi-scale latent variable with a quantization compensation module.

Reference graph

Works this paper leans on

30 extracted references · 12 canonical work pages · cited by 1 Pith paper

[30]

Nature Communications14(1), 4542 (2023)

Zhang, X., Wu, C., Zhang, Y., Xie, W., Wang, Y.: Knowledge-enhanced visual- language pre-training on chest radiology images. Nature Communications14(1), 4542 (2023)

2023
[1]

arxiv 2020

Abnar, S., Zuidema, W.: Quantifying attention flow in transformers. arxiv 2020. arXiv preprint arXiv:2005.0092810(2022)

arXiv 2022
[2]

Medical image anal- ysis66, 101797 (2020)

Bustos, A., Pertusa, A., Salinas, J.M., De La Iglesia-Vaya, M.: Padchest: A large chest x-ray image dataset with multi-label annotated reports. Medical image anal- ysis66, 101797 (2020)

2020
[3]

Nature Machine Intelligence3(7), 610–619 (2021)

DeGrave, A.J., Janizek, J.D., Lee, S.I.: Ai for radiographic covid-19 detection se- lects shortcuts over signal. Nature Machine Intelligence3(7), 610–619 (2021)

2021
[4]

arXiv preprint arXiv:2010.11929 (2020)

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

arXiv 2020
[5]

ACM Computing Surveys56(7), 1–39 (2024)

Gao, Y., Gu, S., Jiang, J., Hong, S.R., Yu, D., Zhao, L.: Going beyond xai: A sys- tematic survey for explanation-guided learning. ACM Computing Surveys56(7), 1–39 (2024)

work page 2024
[6]

In: proceedings of the 28th ACM SIGKDD conference on knowledge discovery and data mining

Gao, Y., Sun, T.S., Bai, G., Gu, S., Hong, S.R., Liang, Z.: Res: A robust frame- work for guiding visual explanation. In: proceedings of the 28th ACM SIGKDD conference on knowledge discovery and data mining. pp. 432–442 (2022)

work page 2022
[7]

Proceedings of the ACM on Human-Computer Interaction6(CSCW2), 1–28 (2022)

Gao, Y., Sun, T.S., Zhao, L., Hong, S.R.: Aligning eyes between humans and deep neural network through interactive attention alignment. Proceedings of the ACM on Human-Computer Interaction6(CSCW2), 1–28 (2022)

work page 2022

Show all 30 references

[8]

Nature Machine In- telligence2(11), 665–673 (2020)

Geirhos, R., Jacobsen, J.H., Michaelis, C., Zemel, R., Brendel, W., Bethge, M., Wichmann, F.A.: Shortcut learning in deep neural networks. Nature Machine In- telligence2(11), 665–673 (2020)

2020
[9]

The Lancet Digital Health 4(6), e406–e414 (2022)

Gichoya, J.W., Banerjee, I., Bhimireddy, A.R., Burns, J.L., Celi, L.A., Chen, L.C., Correa, R., Dullerud, N., Ghassemi, M., Huang, S.C., et al.: Ai recognition of patient race in medical imaging: a modelling study. The Lancet Digital Health 4(6), e406–e414 (2022)

2022
[10]

arXiv preprint arXiv:2110.14755 (2021)

Glocker, B., Jones, C., Bernhardt, M., Winzeck, S.: Algorithmic encoding of pro- tected characteristics in image-based models for disease detection. arXiv preprint arXiv:2110.14755 (2021)

2021 arXiv
[11]

arXiv preprint arXiv:2404.09656 (2024)

Gorbatovski, A., Shaposhnikov, B., Malakhov, A., Surnachev, N., Aksenov, Y., Maksimov, I., Balagansky, N., Gavrilov, D.: Learn your reference model for real good alignment. arXiv preprint arXiv:2404.09656 (2024)

2024 arXiv
[12]

In: Proceedings of the AAAI conference on artificial intelligence

Irvin, J., Rajpurkar, P., Ko, M., Yu, Y., Ciurea-Ilcus, S., Chute, C., Marklund, H., Haghgoo, B., Ball, R., Shpanskaya, K., et al.: Chexpert: A large chest radiograph dataset with uncertainty labels and expert comparison. In: Proceedings of the AAAI conference on artificial in...

2019
[13]

arXiv preprint arXiv:1901.07042 (2019) 10 H

Johnson, A.E., Pollard, T.J., Greenbaum, N.R., Lungren, M.P., Deng, C.y., Peng, Y., Lu, Z., Mark, R.G., Berkowitz, S.J., Horng, S.: Mimic-cxr-jpg, a large publicly available database of labeled chest radiographs. arXiv preprint arXiv:1901.07042 (2019) 10 H. Luo et al

2019 arXiv
[14]

Nature Machine Intelligence6(2), 138–146 (2024)

Jones, C., Castro, D.C., De Sousa Ribeiro, F., Oktay, O., McCradden, M., Glocker, B.: A causal perspective on dataset bias in machine learning for medical imaging. Nature Machine Intelligence6(2), 138–146 (2024)

2024
[15]

Nature Medicine30(4), 1166–1173 (2024)

Ktena, I., Wiles, O., Albuquerque, I., Rebuffi, S.A., Tanno, R., Roy, A.G., Azizi, S., Belgrave, D., Kohli, P., Cemgil, T., et al.: Generative models improve fairness of medical classifiers under distribution shifts. Nature Medicine30(4), 1166–1173 (2024)

2024
[16]

Proceedings of the National Academy of Sciences117(23), 12592–12594 (2020)

Larrazabal, A.J., Nieto, N., Peterson, V., Milone, D.H., Ferrante, E.: Gender im- balance in medical imaging datasets produces biased classifiers for computer-aided diagnosis. Proceedings of the National Academy of Sciences117(23), 12592–12594 (2020)

2020
[17]

Advances in neural information processing systems32(2019)

Lu,J.,Batra,D.,Parikh,D.,Lee,S.:Vilbert:Pretrainingtask-agnosticvisiolinguis- tic representations for vision-and-language tasks. Advances in neural information processing systems32(2019)

2019
[18]

In: International Conference on Medical Image Computing and Computer-Assisted Intervention

Luo, H., de Mortanges, A.P., Inel, O., Reyes, M.: Dwarf: Disease-weighted net- work for attention map refinement. In: International Conference on Medical Image Computing and Computer-Assisted Intervention. pp. 59–68. Springer (2024)

2024
[19]

Scientific Data9(1), 429 (2022)

Nguyen, H.Q., Lam, K., Le, L.T., Pham, H.H., Tran, D.Q., Nguyen, D.B., Le, D.D., Pham, C.M., Tong, H.T., Dinh, D.H., et al.: Vindr-cxr: An open dataset of chest x-rays with radiologist’s annotations. Scientific Data9(1), 429 (2022)

2022
[20]

Science366(6464), 447–453 (2019)

Obermeyer, Z., Powers, B., Vogeli, C., Mullainathan, S.: Dissecting racial bias in an algorithm used to manage the health of populations. Science366(6464), 447–453 (2019)

2019
[21]

In: International conference on machine learning

Rieger, L., Singh, C., Murdoch, W., Yu, B.: Interpretations are useful: penaliz- ing explanations to align neural networks with prior knowledge. In: International conference on machine learning. pp. 8116–8126. PMLR (2020)

2020
[22]

Nature Machine Intelligence4(10), 867– 878 (2022)

Saporta, A., Gui, X., Agrawal, A., Pareek, A., Truong, S.Q., Nguyen, C.D., Ngo, V.D., Seekins, J., Blankenberg, F.G., Ng, A.Y., et al.: Benchmarking saliency methods for chest x-ray interpretation. Nature Machine Intelligence4(10), 867– 878 (2022)

2022
[23]

In: BIOCOMPUTING 2021: proceedings of the Pacific symposium

Seyyed-Kalantari, L., Liu, G., McDermott, M., Chen, I.Y., Ghassemi, M.: Chex- clusion: Fairness gaps in deep chest x-ray classifiers. In: BIOCOMPUTING 2021: proceedings of the Pacific symposium. pp. 232–243. World Scientific (2020)

2020
[24]

Nature medicine27(12), 2176–2182 (2021)

Seyyed-Kalantari, L., Zhang, H., McDermott, M.B., Chen, I.Y., Ghassemi, M.: Underdiagnosisbiasofartificialintelligencealgorithmsappliedtochestradiographs in under-served patient populations. Nature medicine27(12), 2176–2182 (2021)

2021
[25]

Journal of Medical Imaging9(6), 061102–061102 (2022)

Stanley, E.A., Wilms, M., Mouches, P., Forkert, N.D.: Fairness-related performance and explainability effects in deep learning models for brain image analysis. Journal of Medical Imaging9(6), 061102–061102 (2022)

2022
[26]

IEEE Transactions on Medical Imaging41(7), 1688–1698 (2022)

Wang, S., Ouyang, X., Liu, T., Wang, Q., Shen, D.: Follow my eye: Using gaze to supervise computer-aided diagnosis. IEEE Transactions on Medical Imaging41(7), 1688–1698 (2022)

2022
[27]

In: Proceedings of the IEEE CVPR

Wang, X., Peng, Y., Lu, L., Lu, Z., Bagheri, M., Summers, R.M.: Chestx-ray8: Hospital-scale chest x-ray database and benchmarks on weakly-supervised classi- fication and localization of common thorax diseases. In: Proceedings of the IEEE CVPR. pp. 2097–2106 (2017)

2017
[28]

In: International Conference on Medical Image Computing and Computer-Assisted Intervention

Wu, S., Zhang, X., Wang, B., Jin, Z., Li, H., Feng, J.: Gaze-directed vision gnn for mitigating shortcut learning in medical image. In: International Conference on Medical Image Computing and Computer-Assisted Intervention. pp. 514–524. Springer (2024) Aligner 11

2024
[29]

Nature Reviews Methods Primers4(1), 86 (2024)

Zhang, H., Gerych, W., Ghassemi, M.: A data-centric perspective to fair machine learning for healthcare. Nature Reviews Methods Primers4(1), 86 (2024)

2024

Pith tools

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

[1] [30]

Nature Communications14(1), 4542 (2023)

Zhang, X., Wu, C., Zhang, Y., Xie, W., Wang, Y.: Knowledge-enhanced visual- language pre-training on chest radiology images. Nature Communications14(1), 4542 (2023)

2023

[2] [1]

arxiv 2020

Abnar, S., Zuidema, W.: Quantifying attention flow in transformers. arxiv 2020. arXiv preprint arXiv:2005.0092810(2022)

arXiv 2022

[3] [2]

Medical image anal- ysis66, 101797 (2020)

Bustos, A., Pertusa, A., Salinas, J.M., De La Iglesia-Vaya, M.: Padchest: A large chest x-ray image dataset with multi-label annotated reports. Medical image anal- ysis66, 101797 (2020)

2020

[4] [3]

Nature Machine Intelligence3(7), 610–619 (2021)

DeGrave, A.J., Janizek, J.D., Lee, S.I.: Ai for radiographic covid-19 detection se- lects shortcuts over signal. Nature Machine Intelligence3(7), 610–619 (2021)

2021

[5] [4]

arXiv preprint arXiv:2010.11929 (2020)

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

arXiv 2020

[6] [5]

ACM Computing Surveys56(7), 1–39 (2024)

Gao, Y., Gu, S., Jiang, J., Hong, S.R., Yu, D., Zhao, L.: Going beyond xai: A sys- tematic survey for explanation-guided learning. ACM Computing Surveys56(7), 1–39 (2024)

work page 2024

[7] [6]

In: proceedings of the 28th ACM SIGKDD conference on knowledge discovery and data mining

Gao, Y., Sun, T.S., Bai, G., Gu, S., Hong, S.R., Liang, Z.: Res: A robust frame- work for guiding visual explanation. In: proceedings of the 28th ACM SIGKDD conference on knowledge discovery and data mining. pp. 432–442 (2022)

work page 2022

[8] [7]

Proceedings of the ACM on Human-Computer Interaction6(CSCW2), 1–28 (2022)

Gao, Y., Sun, T.S., Zhao, L., Hong, S.R.: Aligning eyes between humans and deep neural network through interactive attention alignment. Proceedings of the ACM on Human-Computer Interaction6(CSCW2), 1–28 (2022)

work page 2022