Pith. sign in

REVIEW 6 major objections 8 minor 48 references

Predictive Representativity: Uncovering Racial Bias in AI-based Skin Cancer Detection

T0 review · 6 major / 8 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read The paper claims that AI-based skin cancer classifiers trained on HAM10000 are systematically less accurate for darker-skinned patients, and it proposes Predictive Representativity, an outcome-level audit metric, to expose such gaps.

desk verdict Useful external-dataset case study showing real precision gaps for darker skin, wrapped in a PR formalism that doesn't hold together; referee it for the empirical result, not the new metric. read the letter →

arxiv 2507.14176 v1 pith:HWTGNSDG submitted 2025-07-10 cs.LG stat.COstat.ML

Andrés Morales-Forero (1) , Lili J. Rueda (2) , Ronald Herrera (3) , Samuel Bassetto (1) , Eric Coatanea (4) ((1) Polytechnique Montréal , (2) Universidad El Bosque , (3) Boehringer Ingelheim International GmbH , (4) Tampere University) This is my paper · ORCID
classification cs.LGstat.COstat.ML
keywords algorithmicfairnessracialbiasinAIPredictiveRepresentativitytransportabilitymedicaloutcomeequityHAM10000Fitzpatrickphototype
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 introduces Predictive Representativity, an audit metric that measures whether a model's predictions align with true outcomes equally across subpopulations, and uses it to test five standard skin-cancer classifiers trained on HAM10000 against an independent Colombian test set labeled with Fitzpatrick skin phototypes. The paper claims that despite proportional sampling in the training data, every architecture underperforms on darker skin, with ResNet50's precision for malignant lesions at 0.897 for lighter skin versus 0.484 for darker skin ($p\leq0.001$). It argues that representativity is a dynamic property of model predictions, not a static feature of dataset composition, and proposes an External Transportability Criterion for deciding when fairness generalizes to a new population. If the paper is right, outcome-level audits are needed to expose subgroup harm that aggregate accuracy and proportional sampling both hide, which matters because darker-skinned melanoma patients already face delayed diagnosis and higher mortality.

What carries the argument

The central object is Predictive Representativity, defined as $PR(P,S,A)=\mathbb{E}_{X\sim P_S}[D(P_S(Y|X)\|\hat{P}_S(Y|X))]-\mathbb{E}_{X\sim P}[D(P(Y|X)\|\hat{P}(Y|X))]$, the difference between the average statistical divergence between true and predicted label distributions within a subpopulation $S$ and the same divergence over the whole population $P$, for a divergence $D$ that the paper leaves open (Kullback-Leibler, Jensen-Shannon, and total variation are listed as candidates). In practice the paper uses metric-level PR, $PR_{M_i}(S)=M_i(S)-M_i(P)$, which compares a domain-relevant performance metric on the subpopulation against its population value; the case study applies this to precision, sensitivity, AUC-PR, specificity, accuracy, AUC-ROC, and F1-score. The External Transportability Criterion declares a model transportable to a new population $P'$ for subgroup $S'$ when $|PR(P',S',A)|\leq\epsilon$. The empirical estimator replaces the theoretical distributions with point masses at the true and predicted labels, and confidence intervals are obtained by nonparametric bootstrap.

What would settle it

The central claim would be overturned if a replication on a larger externally collected test set with Fitzpatrick annotations found no significant malignant-lesion precision gap between lighter (types II–III) and darker (types IV–VI) skin for any of the five architectures trained on HAM10000; the formal quantity would be undermined if the empirical cPR estimator, computed with a specified divergence and bootstrap intervals, failed to track the metric-level subgroup gaps on the same data.

Watch

Extended reading notes

Core claim

The central claim is that AI-based skin cancer classifiers trained on HAM10000 exhibit substantial, statistically significant predictive inequity across skin phototypes when evaluated on an independent clinical test set from Bogotá, Colombia. The discovery is that this inequity is invisible to proportional sampling: darker skin phototypes were included in the training data in proportions roughly consistent with their prevalence, yet all five benchmark CNN architectures — ResNet-50, DenseNet-121, MobileNet-V2, EfficientNet-V2-B0, and VGG-16 — show markedly worse precision, AUC-PR, and F1-score for malignant lesions on darker skin (Fitzpatrick IV–VI) than on lighter skin (II–III). The paper interprets this as evidence that representativity must be redefined as an outcome-level, context-sensitive property, and that Predictive Representativity operationalizes that shift.

Load-bearing premise

The load-bearing premise is that the empirical estimator in Eq. (3), which compares point-mass divergence in a subgroup against the full population, faithfully estimates the theoretical divergence-based Predictive Representativity; the paper does not state which divergence is used, and for Kullback-Leibler divergence that empirical quantity is undefined whenever the predicted and true labels differ, so if the identification fails the case study is measuring ordinary subgroup performance gaps rather than the new PR quantity.

Editorial extensions

If this is right

  • Dataset-level representativeness in the sampling sense does not guarantee equitable conditional predictions, so claims of dataset diversity based on proportional composition alone are insufficient for fairness claims.
  • Any of the five architectures trained on HAM10000 will underperform on darker-skinned patients in a way that is clinically dangerous, because malignant lesions are more likely to be misclassified as benign in that group.
  • Aggregate accuracy of about 0.72 on the external BOSQUE test set coexists with a malignant-lesion precision gap of roughly 0.41 between skin-tone groups, so reporting only overall accuracy hides the subgroup failure.
  • The External Transportability Criterion provides a threshold-based check, $|PR(P',S',A)|\leq\epsilon$, for deciding whether a model's fairness record transfers to a new population.
  • Predictive Representativity is intended as a post-hoc auditing metric rather than a training objective, because directly optimizing for parity across subgroups could overfit to subgroup idiosyncrasies and hurt generalization.

Reading between the lines

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

  • Editorial inference: because metric-level PR is a difference of standard performance metrics, it can be applied to any classification task with labeled subgroups, so the auditing recipe extends beyond dermatology to other regulated domains.
  • Editorial inference: the formal divergence-based definition and the empirical point-mass estimator are not automatically the same quantity unless a specific divergence is chosen and consistency is proven; until then the case-study numbers are best read as ordinary subgroup performance gaps.
  • Editorial inference: a natural next experiment is to audit newer datasets that provide explicit skin-tone labels to see whether deliberate dark-skin curation shrinks the observed PR gaps or merely re-weights lesions the models already handle well.
  • Editorial inference: the paper's decision to exclude synthetic skin darkening can be tested directly by measuring PR after augmentation that preserves real dermatological features, which would show whether the gap closes or new artifacts appear.
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

6 major / 8 minor

Summary. The paper proposes Predictive Representativity (PR), defined in Eq. (1) as the difference between a subpopulation's average divergence from true labels to model predictions and the full population's average divergence, together with an External Transportability Criterion based on a tolerance epsilon. The empirical estimator in Eq. (3) replaces true and predicted labels by point masses. The authors then introduce metric-level PR, PRMi(S)=M_i(S)-M_i(P), and apply it to five CNN classifiers trained on HAM10000 and evaluated on the BOSQUE Test set (n=165) from Colombia, grouped by Fitzpatrick phototype. They report large, statistically significant precision gaps for darker skin (e.g., ResNet50 precision 0.897 light vs 0.484 dark, p<0.001) and conclude that proportional dataset composition does not ensure equitable predictive performance.

Significance. The raw empirical finding is meaningful and clinically important: an independently collected, Fitzpatrick-annotated dataset is used to show a consistent dark-skin precision deficit across five architectures, with small p-values and publicly available code and data. If the formal PR framework were valid, it would offer a practical outcome-level auditing tool. However, as presented, the formal divergence-based PR is not operationalized: Eq. (3) is undefined or trivial for standard divergences, the case study actually reports metric-level differences, and the interpretation of signs is inconsistent with the formal definition. The paper's contribution at present is therefore a conventional subgroup disparity audit plus a proposed but unvalidated formal metric. The empirical evidence supports a call for external validation and dataset documentation, but does not yet establish PR as a new fairness measure.

major comments (6)
  1. [Modelling set-up, Eq. (1)] The formal definition of PR requires S⊆X, a measurable subset of the input space, and defines P_S as the law of (X,Y) given X∈S. In the case study, subgroups are defined by Fitzpatrick phototype, a demographic label that is not part of the dermoscopic image input X; the model does not receive this attribute. The framework therefore does not formally cover the case study's subgroups. The authors should either define a sensitive attribute Z and condition on Z, or justify treating skin phototype as a function of X.
  2. [Empirical Estimation, Eq. (3)] Equation (3) estimates cPR by averaging D(δ_{y_i}∥δ_{f_A(x_i)}). For KL divergence this term is 0 when predictions match and +∞ when they differ; for total variation it is a 0/1 mismatch indicator. The paper does not state which divergence D is used, does not prove consistency of this point-mass estimator for the conditional-divergence quantity in Eq. (1), and does not explain why the model's predictive distribution \hat P(Y|X) should be replaced by a point mass at the hard label. As written, cPR is an error-rate gap, not an estimate of the proposed PR.
  3. [Table 1 and Case Study, Table 2] Table 1 states that PR>0 indicates underperformance and PR<0 indicates overfitting or disproportionate optimization for S. But PRMi(S)=M_i(S)-M_i(P), so a positive PRMi means the subgroup outperforms the population. Table 2 reports PRMi values that are systematically positive for light skin and negative for dark skin, and the text interprets the dark-skin negative values as reduced ability and underperformance (e.g., the F1-Score paragraph). This is the opposite of the formal sign convention. The paper must either redefine PRMi with the opposite sign or correct Table 1; the current inconsistency drives the interpretation of every result.
  4. [Operationalising Predictive Representativity] The metric-level quantity PRMi(S)=M_i(S)-M_i(P) is a standard subgroup performance gap. The paper does not show that PRMi estimates, bounds, or approximates the divergence-based PR of Eq. (1), nor does it connect the thresholds epsilon_i to the divergence scale. Thus the empirical case study does not instantiate the formal PR; it reports a conventional disparity audit under a new name. If the contribution is metric-level PR, the paper should position it as such and relate it to existing fairness metrics.
  5. [External Transportability Criterion and Discussion] The criterion |PR(P',S',A)|≤epsilon is never instantiated: no epsilon value is specified, no source-population PR is computed (HAM10000 lacks skin-tone labels), and the paper does not evaluate |PR| on BOSQUE. The Discussion's claim that models failed to meet a standard PR tolerance is therefore unsupported. Without a specified tolerance and an operational estimator, the criterion is not falsifiable.
  6. [Introduction and Case Study precision paragraph] The paper states that lower precision in darker skin means models are more likely to misclassify cancerous lesions as benign in this group. This is incorrect: precision is TP/(TP+FP), so low precision means more false positives among predicted malignant cases, i.e., benign lesions flagged as malignant. The delay-of-diagnosis harm corresponds to low sensitivity, which the paper finds less affected. The clinical narrative built on the precision result should be corrected.
minor comments (8)
  1. [Case Study, Table 2] The paper reports AUC-PR and AUC-ROC but does not state whether these are computed from softmax scores or from hard predictions; specify the scoring protocol.
  2. [Case Study, Table 2] The Z-statistic's null hypothesis is not stated; clarify whether p-values test PRMi=0 or PRMi(Light)=PRMi(Dark), and whether multiple-comparison corrections were applied across 5 models × 7 metrics.
  3. [Abstract and Case Study] The BOSQUE sample size is stated as 167 in the Abstract and 165 in the Case Study and Table 2; reconcile the discrepancy.
  4. [Introduction and Case Study] The Fitzpatrick grouping is inconsistent: the Introduction refers to lighter types II–III and darker IV–VI, while the Case Study groups I–III as lighter; state which types are actually present in the data.
  5. [Introduction] The claim that all five models exhibited reduced precision and recall rates is contradicted by Table 2 for MobileNetV2 sensitivity (0.899 light vs 0.947 dark); revise to refer to precision only or qualify the recall claim.
  6. [Empirical Estimation, Eq. (3)] The subscript I in S_I is undefined.
  7. [General] The text alternates between representativity and representativeness and contains typos (e.g., benckmark, indentifying meaningfull examples, emploing effective estimators); copyediting is needed.
  8. [Abstract and Introduction] The claim that the finding holds despite proportional sampling in the source data is confusing because the paper also describes HAM10000's darker-skin share as around 5% and disproportionately light; clarify whether proportional refers to source-population prevalence or to the deployment population.

Circularity Check

1 steps flagged · score 4.0 of 10

Metric-level PR is defined as a metric gap; the headline negative PR for darker skin restates that gap by construction, while raw performance disparities stand independently.

  1. self definitional [Operationalising Predictive Representativity (Definition) and Case Study: F1-Score (Malignant)]
    "PRMi (S) =M i(S)− Mi(P) ... ResNet50 achieves an F1-score of 0.892 for lighter skin, compared to 0.600 for darker skin. The PR value is 0.070 for lighter skin and -0.221 for darker skin. The large negative PR for darker skin (Z = 4.386, p≤0.001) reflects the model’s reduced ability to balance precision and recall effectively in this subgroup."

    Definitional reduction: PRMi(S) is literally M_i(S)-M_i(P), so PRMi(Dark) = F1(Dark)-F1(P) = 0.600-0.821 = -0.221 and PRMi(Light) = F1(Light)-F1(P) = 0.070 are arithmetic consequences of the reported F1 values, not independent evidence. The claim that a 'large negative PR for darker skin' indicates reduced ability is a restatement of the fact that dark-skin F1 lies below the population F1; no PR-specific inference is added. The raw metric gap is real and independently measured, but the paper presents PR as a new outcome-level audit while the PR label adds no information beyond the definition of the gap; sign and magnitude follow by construction, and the Z-statistic tests that same gap.

full rationale

The central empirical finding—that DNNs trained on HAM10000 perform substantially worse on darker Fitzpatrick skin types in the BOSQUE test set—is not circular: the raw precision, sensitivity, AUC-PR, specificity, accuracy, AUC-ROC, and F1 values in Table 2 are independent external measurements, and the disparity remains meaningful even if PR is ignored. The circular part is narrower and definitional: metric-level PR is defined as PRMi(S)=M_i(S)-M_i(P), so the reported negative PRMi values for darker skin (e.g., F1: -0.221) and positive values for lighter skin (e.g., F1: 0.070) are exactly the metric gaps by construction. The paper's claim that PR 'reveals' underperformance restates its own definition. The formal divergence-based PR (Eq. 1) and the point-mass estimator (Eq. 3) are never connected to Table 2, which reports metric-level PRMi; this is a rigor/inconsistency concern rather than a circularity. The External Transportability Criterion places a threshold on the same gap with epsilon unspecified, so the statement that classifiers 'failed to meet a standard PR tolerance' has no operational content beyond the criterion's own definition. The one self-citation to Morales-Forero et al. (2024) supports the HAM10000 20:1 lighter-to-darker composition claim, but it is not load-bearing because the BOSQUE evaluation independently establishes the disparity. Overall: partial circularity in the PR presentation, with an independent empirical core.

Assumptions & free parameters 4 free parameters · 5 assumptions · 2 invented entities

The ledger reflects that the paper's formal contribution is a definition, not a derived result. The central empirical comparison depends on domain assumptions about phototype labels, label mapping, and the small external sample. No fitted parameters are used to produce the disparity, but the tolerance epsilon and grouping choices are free choices. PR and the transportability criterion are new named constructs with no external validation.

free parameters (4)
  • epsilon_i tolerance thresholds in metric-level PR and External Transportability Criterion = not specified
    The criterion |PR_{M_i}(S)| <= epsilon_i is vacuous until epsilon_i is set; the paper leaves the choice to stakeholders and gives no guidance or sensitivity analysis.
  • Fitzpatrick light/dark grouping boundary = light: types I-III (n=107); dark: types IV-VI (n=58)
    The main stratified comparison depends on this grouping; the paper does not test alternative groupings or account for the discrepancy between the text (I-III) and Table 2 (II-III).
  • Binary label mapping for malignancy = malignant includes melanoma, basal cell carcinoma, actinic keratoses; benign includes nevi and keratoses
    Actinic keratosis is precancerous, not invasive malignancy; including it changes precision and F1 estimates. The clinical validity of this mapping is not discussed.
  • Oversampling ratio and class weights used in training = not reported
    The paper states oversampling and class weighting were used to address imbalance but does not report the ratios, which influence model behavior and the measured gaps.
assumptions (5)
  • ad hoc to paper The empirical point-mass estimator cPR is a valid estimate of the theoretical divergence-based PR.
    Eq. (3) replaces P(Y|X) and \hat P(Y|X) with point masses delta_{y_i} and delta_{f_A(x_i)}. For KL this is undefined when predicted and true labels differ; no consistency proof is given.
  • domain assumption Fitzpatrick phototype annotations in BOSQUE are accurate and the light/dark grouping captures clinically meaningful skin-tone strata.
    The main stratification uses skin phototype labels assigned by clinical staff; annotation reliability and inter-rater agreement are not reported.
  • domain assumption HAM10000 contains roughly 5% darker skin phototypes (20:1 ratio), cited from the authors' prior paper.
    The argument that HAM10000 is proportionally representative relies on Morales-Forero et al. (2024); no independent estimate is provided.
  • domain assumption Standard CNN architectures trained with binary labels and augmentations are representative of AI-based skin cancer classifiers.
    Only five ImageNet-style CNNs are evaluated; no segmentation models, foundation models, or dermatology-specific models are included.
  • ad hoc to paper The divergence D exists and is finite for the distributions considered.
    PR is defined for a generic divergence D, but the experiments do not state which D is used, so the theoretical statements lack a concrete mathematical object.
invented entities (2)
  • Predictive Representativity (PR) divergence gap
    purpose: Quantify subpopulation-specific predictive misalignment relative to the population average.
    PR is defined in the paper and is not anchored to an external benchmark; in its empirical form it reduces to standard error-rate gaps.
  • External Transportability Criterion (|PR| <= epsilon)
    purpose: Declare a model fairly transportable to a new population if the PR gap is below a stakeholder-defined threshold.
    The criterion is a definition; the threshold epsilon is not specified or validated against known transportable and non-transportable models.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Predictive Representativity: Uncovering Racial Bias in AI-based Skin Cancer Detection." pith.science (2026). https://pith.science/paper/HWTGNSDG

@misc{pith2026250714176,
  author       = {Pith},
  title        = {Pith review of: Predictive Representativity: Uncovering Racial Bias in AI-based Skin Cancer Detection},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HWTGNSDG}},
  note         = {Machine review of arXiv:2507.14176}
}
read the original abstract

Artificial intelligence (AI) systems increasingly inform medical decision-making, yet concerns about algorithmic bias and inequitable outcomes persist, particularly for historically marginalized populations. This paper introduces the concept of Predictive Representativity (PR), a framework of fairness auditing that shifts the focus from the composition of the data set to outcomes-level equity. Through a case study in dermatology, we evaluated AI-based skin cancer classifiers trained on the widely used HAM10000 dataset and on an independent clinical dataset (BOSQUE Test set) from Colombia. Our analysis reveals substantial performance disparities by skin phototype, with classifiers consistently underperforming for individuals with darker skin, despite proportional sampling in the source data. We argue that representativity must be understood not as a static feature of datasets but as a dynamic, context-sensitive property of model predictions. PR operationalizes this shift by quantifying how reliably models generalize fairness across subpopulations and deployment contexts. We further propose an External Transportability Criterion that formalizes the thresholds for fairness generalization. Our findings highlight the ethical imperative for post-hoc fairness auditing, transparency in dataset documentation, and inclusive model validation pipelines. This work offers a scalable tool for diagnosing structural inequities in AI systems, contributing to discussions on equity, interpretability, and data justice and fostering a critical re-evaluation of fairness in data-driven healthcare.

Figures

Figures reproduced from arXiv: 2507.14176 by the authors.

Figure 1
Figure 1. Predictive Representativity and External Transportability Criterion. The model is trained and [PITH_FULL_IMAGE:figures/full_fig_p010_1.png] view at source ↗
Figure 2
Figure 2. Predictive Representativity metrics by model and skin type group. [PITH_FULL_IMAGE:figures/full_fig_p014_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

48 extracted references · 47 canonical work pages

  1. [1]

    , Hardt, M

    barocas2023fairness APACrefauthors Barocas, S. , Hardt, M. \ Narayanan, A. APACrefauthors \ 2023 . Fairness and machine learning: Limitations and opportunities Fairness and machine learning: Limitations and opportunities . MIT press

    work page 2023

  2. [2]

    \ Friedman, B

    bender2018data APACrefauthors Bender, E M. \ Friedman, B. APACrefauthors \ 2018 . Data statements for natural language processing: Toward mitigating system bias and enabling better science Data statements for natural language processing: Toward mitigating system bias and enabling better science . Transactions of the Association for Computational Linguisti...

    work page 2018

  3. [3]

    , Kashlan, R

    brady2021racial APACrefauthors Brady, J. , Kashlan, R. , Ruterbusch, J. , Farshchian, M. \ Moossavi, M. APACrefauthors \ 2021 . Racial disparities in patients with melanoma: a multivariate survival analysis Racial disparities in patients with melanoma: a multivariate survival analysis . Clinical, Cosmetic and Investigational Dermatology 547--550

    work page 2021

  4. [4]

    , Bowyer, K W

    chawla2002smote APACrefauthors Chawla, N V. , Bowyer, K W. , Hall, L O. \ Kegelmeyer, W P. APACrefauthors \ 2002 . SMOTE: synthetic minority over-sampling technique Smote: synthetic minority over-sampling technique . Journal of artificial intelligence research 16 321--357

    work page 2002

  5. [5]

    \ Kj rsgaard, R D

    clemmensen2022data APACrefauthors Clemmensen, L H. \ Kj rsgaard, R D. APACrefauthors \ 2022 . Data representativity for machine learning and AI systems Data representativity for machine learning and ai systems . arXiv preprint arXiv:2203.04706

    arXiv 2022

  6. [6]

    , Pierson, E

    corbett2017algorithmic APACrefauthors Corbett-Davies, S. , Pierson, E. , Feller, A. , Goel, S. \ Huq, A. APACrefauthors \ 2017 . Algorithmic decision making and the cost of fairness Algorithmic decision making and the cost of fairness . Proceedings of the 23rd acm sigkdd international conference on knowledge discovery and data mining Proceedings of the 23...

    work page 2017

  7. [7]

    , Vasquez, J

    cortez2021impact APACrefauthors Cortez, J L. , Vasquez, J. \ Wei, M L. APACrefauthors \ 2021 . The impact of demographics, socioeconomics, and health care access on melanoma outcomes The impact of demographics, socioeconomics, and health care access on melanoma outcomes . Journal of the American Academy of Dermatology 84 6 1677--1683

    work page 2021

  8. [8]

    \ Lunsford, N B

    culp2019peer APACrefauthors Culp, M B. \ Lunsford, N B. APACrefauthors \ 2019 . Peer reviewed: Melanoma among non-Hispanic black Americans Peer reviewed: Melanoma among non-hispanic black americans . Preventing Chronic Disease 16

    work page 2019

Show all 48 references
  1. [9]

    \ Davey Smith, G

    ebrahim2013commentary APACrefauthors Ebrahim, S. \ Davey Smith, G. APACrefauthors \ 2013 . Commentary: Should we always deliberately be non-representative? Commentary: Should we always deliberately be non-representative? International journal of epidemiology 42 4 1022--1026

    2013

  2. [10]

    APACrefauthors \ 2013

    elwood2013commentary APACrefauthors Elwood, J M. APACrefauthors \ 2013 . Commentary: on representativeness Commentary: on representativeness . International journal of epidemiology 42 4 1014--1015

    2013

  3. [11]

    APACrefauthors \ 2020

    epstein2020color APACrefauthors Epstein, H. APACrefauthors \ 2020 . Why the Color of Your Skin Can Affect the Quality of Your Diagnosis Why the color of your skin can affect the quality of your diagnosis . Society to Improve Diagnosis in Medicine

    2020

  4. [12]

    , Friedler, S A

    feldman2015certifying APACrefauthors Feldman, M. , Friedler, S A. , Moeller, J. , Scheidegger, C. \ Venkatasubramanian, S. APACrefauthors \ 2015 . Certifying and removing disparate impact Certifying and removing disparate impact . proceedings of the 21th ACM SIGKDD internation...

    2015

  5. [13]

    , Morgenstern, J

    gebru2021datasheets APACrefauthors Gebru, T. , Morgenstern, J. , Vecchione, B. , Vaughan, J W. , Wallach, H. , Iii, H D. \ Crawford, K. APACrefauthors \ 2021 . Datasheets for datasets Datasheets for datasets . Communications of the ACM 64 12 86--92

    2021

  6. [14]

    APACrefauthors \ 2008

    gohara2008skin APACrefauthors Gohara, M A. APACrefauthors \ 2008 . Skin cancer in skins of color. Skin cancer in skins of color. Journal of drugs in dermatology: JDD 7 5 441--445

    2008

  7. [15]

    , Pouget-Abadie, J

    goodfellow2014generative APACrefauthors Goodfellow, I. , Pouget-Abadie, J. , Mirza, M. , Xu, B. , Warde-Farley, D. , Ozair, S. Bengio, Y. APACrefauthors \ 2014 . Generative adversarial nets Generative adversarial nets . Advances in neural information processing systems 27

    2014

  8. [16]

    \ Schelin, L

    grafstrom2014select APACrefauthors Grafstr \"o m, A. \ Schelin, L. APACrefauthors \ 2014 . How to select representative samples How to select representative samples . Scandinavian Journal of Statistics 41 2 277--290

    2014

  9. [17]

    , Bharadwaj, M

    gupta2016skin APACrefauthors Gupta, A K. , Bharadwaj, M. \ Mehrotra, R. APACrefauthors \ 2016 . Skin cancer concerns in people of color: risk factors and prevention Skin cancer concerns in people of color: risk factors and prevention . Asian Pacific journal of cancer preventio...

    2016

  10. [18]

    , Price, E

    hardt2016equality APACrefauthors Hardt, M. , Price, E. \ Srebro, N. APACrefauthors \ 2016 . Equality of opportunity in supervised learning Equality of opportunity in supervised learning . Advances in neural information processing systems 29

    2016

  11. [19]

    , Sanchez Zapata, M J

    DVN/AQEPIN_2025 APACrefauthors Jaramillo Arboleda, A. , Sanchez Zapata, M J. , RUEDA JAIME, L J. , Morales-Forero, A. \ Bassetto, S. APACrefauthors \ 2025 . BOSQUE Test set . BOSQUE Test set . Harvard Dataverse . APACrefURL https://doi.org/10.7910/DVN/AQEPIN APACrefURL APACref...

    2025 doi

  12. [20]

    , Boone, S L

    kim2009perception APACrefauthors Kim, M. , Boone, S L. , West, D P. , Rademaker, A W. , Liu, D. \ Kundu, R V. APACrefauthors \ 2009 . Perception of skin cancer risk by those with ethnic skin Perception of skin cancer risk by those with ethnic skin . Archives of dermatology 145...

    2009

  13. [21]

    APACrefauthors \ 1965

    kish1965survey APACrefauthors Kish, L. APACrefauthors \ 1965 . Survey sampling. Survey sampling. Wiley

    1965

  14. [22]

    \ Mosteller, F

    kruskal1979III APACrefauthors Kruskal, W. \ Mosteller, F. APACrefauthors \ 1979 1 . Representative sampling, III: The current statistical literature Representative sampling, iii: The current statistical literature . International Statistical Review/Revue Internationale de Stat...

    1979

  15. [23]

    \ Mosteller, F

    kruskal1979II APACrefauthors Kruskal, W. \ Mosteller, F. APACrefauthors \ 1979 2 . Representative sampling, II: Scientific literature, excluding statistics Representative sampling, ii: Scientific literature, excluding statistics . International Statistical Review/Revue Interna...

    1979

  16. [24]

    \ Mosteller, F

    kruskal1979I APACrefauthors Kruskal, W. \ Mosteller, F. APACrefauthors \ 1979 3 . Representative sampling, I: Non-scientific literature Representative sampling, i: Non-scientific literature . International Statistical Review/Revue Internationale de Statistique 13--24

    1979

  17. [25]

    \ Mosteller, F

    kruskal1980representativeIV APACrefauthors Kruskal, W. \ Mosteller, F. APACrefauthors \ 1980 . Representative sampling, IV: The history of the concept in statistics, 1895-1939 Representative sampling, iv: The history of the concept in statistics, 1895-1939 . International Stat...

    1980

  18. [26]

    APACrefauthors \ 2008

    lavrakas2008encyclopedia APACrefauthors Lavrakas, P J. APACrefauthors \ 2008 . Encyclopedia of survey research methods Encyclopedia of survey research methods . Sage publications

    2008

  19. [27]

    , Jia, J

    lester2020absence APACrefauthors Lester, J. , Jia, J. , Zhang, L. , Okoye, G. \ Linos, E. APACrefauthors \ 2020 . Absence of images of skin of colour in publications of COVID-19 skin manifestations Absence of images of skin of colour in publications of covid-19 skin manifestat...

    2020

  20. [28]

    \ Wilkes, R

    louie2018representations APACrefauthors Louie, P. \ Wilkes, R. APACrefauthors \ 2018 . Representations of race and skin tone in medical textbook imagery Representations of race and skin tone in medical textbook imagery . Social Science & Medicine 202 38--42

    2018

  21. [29]

    APACrefauthors \ 1943

    maslow1943theory APACrefauthors Maslow, A H. APACrefauthors \ 1943 . A theory of human motivation. A theory of human motivation. Psychological review 50 4 370

    1943

  22. [30]

    , Morstatter, F

    mehrabi2021survey APACrefauthors Mehrabi, N. , Morstatter, F. , Saxena, N. , Lerman, K. \ Galstyan, A. APACrefauthors \ 2021 . A survey on bias and fairness in machine learning A survey on bias and fairness in machine learning . ACM computing surveys (CSUR) 54 6 1--35

    2021

  23. [31]

    APACrefauthors \ 1962

    merleau1962phenomenology APACrefauthors Merleau-Ponty, M. APACrefauthors \ 1962 . Phenomenology of Perception Routledge & Kegan Paul Phenomenology of perception routledge & kegan paul . London, UK 182

    1962

  24. [32]

    , Subramanian, M

    merrill2016worldwide APACrefauthors Merrill, S J. , Subramanian, M. \ Godar, D E. APACrefauthors \ 2016 . Worldwide cutaneous malignant melanoma incidences analyzed by sex, age, and skin type over time (1955--2007): Is HPV infection of androgenic hair follicular melanocytes a ...

    2016

  25. [33]

    mitchell2019model APACrefauthors Mitchell, M. , Wu, S. , Zaldivar, A. , Barnes, P. , Vasserman, L. , Hutchinson, B. Gebru, T. APACrefauthors \ 2019 . Model cards for model reporting Model cards for model reporting . Proceedings of the conference on fairness, accountability, an...

    2019

  26. [34]

    , Thakral, K

    mittal2024responsible APACrefauthors Mittal, S. , Thakral, K. , Singh, R. , Vatsa, M. , Glaser, T. , Canton Ferrer, C. \ Hassner, T. APACrefauthors \ 2024 . On responsible machine learning datasets emphasizing fairness, privacy and regulatory norms with examples in biometrics ...

    2024

  27. [35]

    , Rueda Jaime, L

    morales2024insight APACrefauthors Morales-Forero, A. , Rueda Jaime, L. , Gil-Qui \ n ones, S R. , Barrera Monta \ n ez, M Y. , Bassetto, S. \ Coatanea, E. APACrefauthors \ 2024 . An insight into racial bias in dermoscopy repositories: A HAM10000 data set analysis An insight in...

    2024

  28. [36]

    \ Bareinboim, E

    pearl2011transportability APACrefauthors Pearl, J. \ Bareinboim, E. APACrefauthors \ 2011 . Transportability of causal and statistical relations: A formal approach Transportability of causal and statistical relations: A formal approach . Proceedings of the AAAI Conference on A...

    2011

  29. [37]

    \ Berton, L

    rabonato2024systematic APACrefauthors Rabonato, R T. \ Berton, L. APACrefauthors \ 2024 . A systematic review of fairness in machine learning A systematic review of fairness in machine learning . AI and Ethics 1--12

    2024

  30. [38]

    , Pizzi, C

    richiardi2013commentary APACrefauthors Richiardi, L. , Pizzi, C. \ Pearce, N. APACrefauthors \ 2013 . Commentary: Representativeness is usually not necessary and often should be avoided Commentary: Representativeness is usually not necessary and often should be avoided . Inter...

    2013

  31. [39]

    , Kunder, V

    rizvi2022bias APACrefauthors Rizvi, Z. , Kunder, V. , Stewart, H. , Torres, P. , Moon, S. , Lingappa, N. others APACrefauthors \ 2022 . The bias of physicians and lack of education in patients of color with melanoma as causes of increased mortality: a scoping review The bias o...

    2022

  32. [40]

    rogers1959theory APACrefauthors Rogers, C R. \ . APACrefauthors \ 1959 . A theory of therapy, personality, and interpersonal relationships: As developed in the client-centered framework A theory of therapy, personality, and interpersonal relationships: As developed in the clie...

    1959

  33. [41]

    , Gallacher, J E

    rothman2013rebuttal APACrefauthors Rothman, K J. , Gallacher, J E. \ Hatch, E E. APACrefauthors \ 2013 1 . Rebuttal: When it comes to scientific inference, sometimes a cigar is just a cigar Rebuttal: When it comes to scientific inference, sometimes a cigar is just a cigar . In...

    2013

  34. [42]

    , Gallacher, J E

    rothman2013representativeness APACrefauthors Rothman, K J. , Gallacher, J E. \ Hatch, E E. APACrefauthors \ 2013 2 . Why representativeness should be avoided Why representativeness should be avoided . International journal of epidemiology 42 4 1012--1014

    2013

  35. [43]

    , Zhong, Y

    rudolph2023defining APACrefauthors Rudolph, J E. , Zhong, Y. , Duggal, P. , Mehta, S H. \ Lau, B. APACrefauthors \ 2023 . Defining representativeness of study samples in medical and population health research Defining representativeness of study samples in medical and populati...

    2023

  36. [44]

    APACrefauthors \ 2012

    swanson2012uk APACrefauthors Swanson, J M. APACrefauthors \ 2012 . The UK Biobank and selection bias The uk biobank and selection bias . The Lancet 380 9837 110

    2012

  37. [45]

    APACrefauthors \ 2006

    tille2006sampling APACrefauthors Till \'e , Y. APACrefauthors \ 2006 . Sampling algorithms Sampling algorithms . Springer

    2006

  38. [46]

    , Rosendahl, C

    tschandl2018ham10000 APACrefauthors Tschandl, P. , Rosendahl, C. \ Kittler, H. APACrefauthors \ 2018 . The HAM10000 dataset, a large collection of multi-source dermatoscopic images of common pigmented skin lesions The ham10000 dataset, a large collection of multi-source dermat...

    2018

  39. [47]

    \ Kahneman, D

    tversky1974judgment APACrefauthors Tversky, A. \ Kahneman, D. APACrefauthors \ 1974 . Judgment under Uncertainty: Heuristics and Biases: Biases in judgments reveal some heuristics of thinking under uncertainty. Judgment under uncertainty: Heuristics and biases: Biases in judgm...

    1974

  40. [48]

    , Eide, M J

    wu2011racial APACrefauthors Wu, X C. , Eide, M J. , King, J. , Saraiya, M. , Huang, Y. , Wiggins, C. others APACrefauthors \ 2011 . Racial and ethnic variations in incidence and survival of cutaneous melanoma in the United States, 1999-2006 Racial and ethnic variations in inci...

    2011

Pith tools

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