Pith. sign in

REVIEW 4 major objections 5 minor 64 references

FairUDT: Fairness-aware Uplift Decision Trees

T0 review · 4 major / 5 minor · reviewed 2026-08-09 · deepseek-v4-flash

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

Pith's one-line read FairUDT claims that a decision tree grown to maximize divergence between favored and deprived groups' class probabilities can pinpoint discriminatory subgroups, and that relabeling only those leaves removes bias while keeping accuracy.

desk verdict Real idea, invalid headline result: the AOD=0.00 is an artifact of relabeling the test set, while the raw-test-set numbers are decent but not state-of-the-art. read the letter →

arxiv 2502.01188 v1 pith:2ONRKS2J submitted 2025-02-03 cs.LG stat.ML

Anam Zahid , Abdur Rehman Ali , Shaina Raza , Rai Shahnawaz , Faisal Kamiran , Asim Karim This is my paper · ORCID
classification cs.LGstat.ML MSC 68T0562H30
keywords algorithmicfairnessdiscriminationidentificationupliftmodelingdecisiontreesleafrelabelingdemographicparityaverageoddsdifferenceinterpretableAI
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

FairUDT claims that bias can be located by treating the favored group as a treatment arm and the deprived group as a control arm, then growing a decision tree whose splits maximize the divergence between the two groups' class-probability distributions. The leaves that end up with a large favored-minus-deprived probability gap are flagged as discriminatory regions, and a relabeling step flips only records in those leaves until the gap closes. Because relabeling is selective rather than whole-group, the paper argues, fairness improves without the usual collapse in accuracy. On Adult, COMPAS, and German Credit used as preprocessed inputs to standard classifiers, the method reports a perfect average odds difference on the relabeled Adult test set and state-of-the-art demographic parity and average odds difference on COMPAS and German Credit. The same trees double as an interpretable audit tool: the leaf paths spell out which attribute combinations form the biased subgroups.

What carries the argument

The engine is the divergence gain $D_{\mathrm{gain}}(A) = D(P^F(Y):P^D(Y)|A) - D(P^F(Y):P^D(Y))$, where $D$ is either KL divergence or squared Euclidean distance, normalized by a split-information term that penalizes splits separating the favored and deprived groups into different subtrees and splits with many outcomes. KL divergence is directional, so it reports how much the favored distribution diverges from the deprived reference distribution, which the paper argues is exactly the direction of preferential treatment. The companion mechanism is leaf relabeling: only leaves with $discl > \sigma_t$ are edited, and within those leaves only enough randomly chosen records are promoted or demoted to force $P^F(y+|l)=P^D(y+|l)$ and $P^F(y-|l)=P^D(y-|l)$; the threshold $\sigma_t \in [0,2]$ is the accuracy-fairness dial.

What would settle it

Run FairUDT on synthetic data with equal ground-truth outcomes across groups but a strong confounder correlated with the sensitive attribute; if the tree flags leaves and relabels them, the identification step has detected association, not discrimination, falsifying the central claim for that setting.

Watch

Extended reading notes

Core claim

The paper's central claim is that uplift modeling can be repurposed from marketing to discrimination identification: when the favored group is treated as treatment and the deprived group as control, the divergence $D(P^F(Y):P^D(Y)|A)$ before and after a split measures how much a feature separates the groups' outcomes, and maximizing that divergence grows a tree whose leaves are precisely the subgroups where favoritism concentrates. At each leaf the paper defines discrimination as $discl = (P^F(y+|l)-P^D(y+|l)) + (P^D(y-|l)-P^F(y-|l))$, relabeling a leaf only when this exceeds a tunable threshold $\sigma_t$, by promoting deprived negative instances when the leaf's majority class is positive and demoting favored positive instances when it is negative. The paper reports that this procedure, run with KL-divergence splitting and logistic regression, reaches AOD 0.00 on the relabeled Adult test set and state-of-the-art DP and AOD on COMPAS and German Credit, with accuracy equal to or better than raw data in those cases. It also claims the resulting trees are sparser and shallower than ordinary decision trees, making the detected discriminatory subgroups directly readable as attribute rules.

Load-bearing premise

The paper assumes that a class-probability gap between favored and deprived groups at a leaf is discrimination; if that gap comes from confounding or real group differences rather than unfair treatment, relabeling removes signal instead of bias.

Editorial extensions

If this is right

  • Used as a pre-processor, FairUDT lets any downstream classifier inherit lower group disparity without changing the classifier, since the relabeled dataset is what is trained on.
  • On COMPAS and German Credit, FairUDT beats the compared pre-processing baselines on DP and AOD while keeping balanced accuracy at or near the raw-data level, so the tradeoff is acceptable rather than prohibitive.
  • Because relabeling targets only leaves above the threshold, it preserves more of the original labels than whole-group relabeling, which is why the paper reports raw-level accuracy on Adult and a 1% accuracy gain on German Credit.
  • The interpretability result means the method can be used as an audit tool: each discriminatory leaf corresponds to an if-then subgroup that a human can read and verify.
  • The tunable threshold $\sigma_t$ gives practitioners a single knob to dial between fairness and accuracy at subgroup level rather than across the whole dataset.

Reading between the lines

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

  • If the favored/deprived gap is not caused by preferential treatment but by an unmeasured confounder or by genuine differences in qualifications, the leaf relabeling will edit labels that are not biased; a causal or counterfactual check on the identified leaves would be needed before the method's 'discrimination' claim is used for real decisions.
  • A synthetic-data test with known ground-truth bias could separate detection from confounding: inject a gap only through a confounder and see whether FairUDT relabels; if it does, the divergence criterion is measuring association, not discrimination.
  • The same divergence machinery could be extended to multi-valued sensitive attributes by replacing the binary favored/deprived split with multiple treatment arms, which the paper lists as future work but does not develop.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper proposes FairUDT, a decision-tree-based data pre-processor that uses divergence between favored and deprived class distributions to grow a tree, identifies leaves with a positive discrimination score discl, and then relabels selected records so that favored and deprived positive/negative rates match in those leaves. The pre-processed data are used to train classifiers, and the method is evaluated on Adult, COMPAS, and German Credit against pre-processing baselines using DP, AOD, BA, and accuracy. The paper claims state-of-the-art fairness results, including AOD 0.00 on the relabeled Adult test set, as well as better interpretability than standard decision trees.

Significance. The idea of adapting uplift modeling to discrimination identification is potentially interesting, and the paper provides a concrete tree-splitting framework with KL and Euclidean divergence gains, a proposed leaf relabeling procedure, and a human-subject interpretability comparison. The code is made available, which is a strength. However, the empirical claims do not currently survive scrutiny: the headline fairness numbers are computed on test labels that the method itself has altered, the relabeling algorithm as printed contains a dimensional error in its count formulas, and the favored/deprived comparison is treated as causal discrimination without supporting assumptions. These issues are load-bearing for the paper's central claim of an acceptable, state-of-the-art accuracy-discrimination tradeoff, so the current evidence does not justify acceptance.

major comments (4)
  1. [§4, §6.1, Table 3] The 'FairUDT + Relabelled Test Set' rows in Table 3 are not a valid evaluation of discrimination removal. Section 4 states that 'the test data must also be pre-processed before making predictions with the classifier,' and Algorithm 1 is designed to enforce P_F(y+|l)=P_D(y+|l) and P_F(y-|l)=P_D(y-|l) in selected leaves. AOD computed on these modified test labels therefore measures how well the classifier predicts labels that the method has already pushed toward group parity, not how much discrimination remains in the original data. The reported AOD of 0.00 for Adult is an artifact of this procedure, and it is not comparable to the AOD values reported for Disparate Impact Remover, Optimized Pre-processing, or Reweighing, which are evaluated on original test labels. The only comparable evidence is the 'FairUDT + Raw Test Set' rows (Adult AOD 0.04, COMPAS 0.02, German 0.02), and those numbers do not by themselves establish the claimed state-of-the-art tradeoff.
  2. [Algorithm 1, lines 13 and 20] The formulas for the number of records to relabel mix probabilities with counts. In Promote, p is computed as floor(P(y+|sF)P(sD) - P(y+|sD)), and in Demote as floor(P(y-|sD)P(sF) - P(y-|sF)). These expressions have probability units, not count units, and they do not include the leaf-specific number of deprived or favored records. To impose P_F(y+|l)=P_D(y+|l), the number of deprived negative records to promote should be N_D^l(P_F(y+|l)-P_D(y+|l)) (and analogously for demotion), with an additional check that this number does not exceed the available records in the leaf. As printed, the algorithm cannot be implemented as intended and is not reproducible.
  3. [§2, §3.4, Eq. (3.11)] The paper defines discrimination as the class-probability difference between favored and deprived groups sharing the same characteristics, and then uses that criterion to relabel data. This is a valid statistical parity criterion only if the favored/deprired split is treated as a randomized or conditionally ignorable treatment; the paper does not state or test this causal assumption, nor does it control for unmeasured confounders. As written, the method can remove any disparity correlated with the sensitive attribute, including disparities that reflect legitimate qualification differences. The central claim that FairUDT removes discrimination rather than group disparity therefore needs either an explicit causal identification argument or a narrower, clearly stated statistical-disparity interpretation.
  4. [§6.1, Figure 3] The discrimination threshold sigma_t is tuned per dataset to the reported optimum (Adult 0.61, COMPAS 0.1, German 1.64), and the paper states that these are the values used for the final results. Because the same test sets are used for tuning and for reporting, the best DP/AOD numbers in Table 3 are selected values rather than predictive estimates. Table 3 also reports no confidence intervals or significance tests. To establish the claimed tradeoff, the authors should report results on a validation split, or show the full sigma_t sweep with confidence bands and a clear separation between tuning and evaluation.
minor comments (5)
  1. [Table 2] The dataset name 'COMP AS' appears to be a typo for 'COMPAS'.
  2. [Appendix B, Proposition 3.2 proof] The sentence replacing dy(a) with 'dy (class probabilities test independence for favored distribution)' should refer to the deprived distribution, not the favored one.
  3. [§3.4, Algorithm 1 line 6] Definition 2 says a leaf is relabeled only if discl > sigma_t, while Algorithm 1 uses disc(l) >= sigma_t; the inconsistency should be resolved, especially for sigma_t = 0.
  4. [§5.2] The relationship between the 75-25 train-test split and the 10-fold cross-validation is not clear; the paper should state whether cross-validation is applied within the training portion and how the test set is used across folds.
  5. [Figure 2] The text says AUC is lower after pre-processing, but Figure 2 only shows ROC curves and no AUC values are reported; the numerical AUC values should be given.

Circularity Check

2 steps flagged · score 7.0 of 10

Relabeled-test-set AOD=0.00 and threshold-tuned state-of-the-art values are evaluation artifacts; raw test-set results retain independent content.

  1. self definitional [Section 4 (test-set pre-processing), Algorithm 1, Table 3 and Section 6.1]
    "It is important to note that the test data must also be pre-processed before making predictions with the classifier. [...] such that the equality conditions P (y+|sF ) = P (y+|sD) and P (y−|sF ) = P (y−|sD) are satisfied in this subgroup. [...] For the Adult dataset, FairUDT achieves a perfect AOD of 0.00 on its relabeled test set with a discrimination threshold of σt = 0.61."

    Algorithm 1 relabels instances in discriminatory leaves until P_F(y+|l)=P_D(y+|l) and P_F(y-|l)=P_D(y-|l), i.e., the test labels used as ground truth are themselves modified to have group parity at leaf level. The AOD=0.00 in the 'FairUDT + Relabelled Test Set' row is therefore measured against a target that the method has parity-adjusted, not against the original outcomes; it shows how well a classifier reproduces the method's own relabeling, not how much discrimination remains in the original data. It also cannot be compared with the Raw, Disparate Impact Remover, Optimized Pre-processing, or Reweighing rows, which use original labels. The 'state-of-the-art' claim built on this perfect AOD is thus supported by construction rather than by independent evaluation.

  2. fitted input called prediction [Section 6.1, Table 3, Figure 3 (σt threshold tuning)]
    "In Table 3, all the datasets are pre-processed using the optimal discrimination thresholds of σt. As shown, the optimal value of 0.00 discrimination for these metrics can be achieved at certain thresholds of σt."

    The threshold σt is not fixed a priori; it is tuned per dataset until the reported DP/AOD reaches the optimum (0.00 or near-0.00), as Figure 3 explicitly demonstrates. The reported 'state-of-the-art' fairness metrics are thus selected by sweeping the threshold on the evaluation data, making the headline numbers a fit to the fairness metric rather than an out-of-sample prediction. This applies to both the raw and relabeled test-set rows, although only the relabeled rows additionally suffer from the label-construction artifact.

full rationale

FairUDT's KL/Egain splitting criteria are not circular: they are adapted from the uplift-modeling divergences of Rzepakowski and Jaroszewicz and are evaluated on unmodified labels in the raw-test-set rows. Those raw rows contain genuine independent results, most notably COMPAS DP=0.00 and the best raw-test AOD values on COMPAS and German Credit. The circularity is concentrated in the paper's headline 'perfect AOD of 0.00' and the threshold-tuned 'state-of-the-art' comparisons. The relabeled test set is produced by Algorithm 1 to satisfy the same leaf-level group-parity conditions used in the fairness target, so an AOD computed against those labels is partly a self-constructed number. The additional per-dataset selection of σt to reach the reported optima turns the headline numbers into a fit. The causal assumption underlying 'class probabilities difference' is a scientific limitation, not a circularity, so it is not scored here. Overall this is a partial circularity: some claims reduce by construction, while the raw-test-set results and the tree-building methodology retain independent content. Score: 7.

Assumptions & free parameters 1 free parameters · 4 assumptions · 0 invented entities

The only fitted parameter is sigma_t, tuned per dataset. The method also assumes the favored/deprived comparison identifies discrimination and that relabeling test labels is legitimate; both are unstated causal or methodological assumptions.

free parameters (1)
  • discrimination threshold sigma_t = Adult 0.61, COMPAS 0.1, German Credit 1.64
    Selected per dataset to reach near-zero DP/AOD; reported results use these optimal values.
assumptions (4)
  • domain assumption A class-probability difference between favored and deprived groups within a leaf indicates discrimination.
    Introduced in Section 2 and Definition 1; treats association as bias without causal identification.
  • domain assumption Favored and deprived groups can be treated as treatment and control arms of an uplift study.
    Section 3.1 and 3.3 use uplift terminology; no randomization or confounding control is provided.
  • ad hoc to paper Relabeling test-set labels is a valid evaluation procedure.
    Section 5 states test data must also be pre-processed; Table 3 reports metrics on the relabeled test set.
  • standard math KL divergence is non-negative (Gibbs inequality) and equals zero only for identical distributions.
    Used in Propositions 3.1 and 3.2 in Appendix B.

how reviews work

0 comments
Cite this review

Pith. "Pith review of FairUDT: Fairness-aware Uplift Decision Trees." pith.science (2026). https://pith.science/paper/2ONRKS2J

@misc{pith2026250201188,
  author       = {Pith},
  title        = {Pith review of: FairUDT: Fairness-aware Uplift Decision Trees},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2ONRKS2J}},
  note         = {Machine review of arXiv:2502.01188}
}
read the original abstract

Training data used for developing machine learning classifiers can exhibit biases against specific protected attributes. Such biases typically originate from historical discrimination or certain underlying patterns that disproportionately under-represent minority groups, such as those identified by their gender, religion, or race. In this paper, we propose a novel approach, FairUDT, a fairness-aware Uplift-based Decision Tree for discrimination identification. FairUDT demonstrates how the integration of uplift modeling with decision trees can be adapted to include fair splitting criteria. Additionally, we introduce a modified leaf relabeling approach for removing discrimination. We divide our dataset into favored and deprived groups based on a binary sensitive attribute, with the favored dataset serving as the treatment group and the deprived dataset as the control group. By applying FairUDT and our leaf relabeling approach to preprocess three benchmark datasets, we achieve an acceptable accuracy-discrimination tradeoff. We also show that FairUDT is inherently interpretable and can be utilized in discrimination detection tasks. The code for this project is available https://github.com/ara-25/FairUDT

Figures

Figures reproduced from arXiv: 2502.01188 by the authors.

Figure 1
Figure 1. End-to-end pipeline of FairUDT and leaf relabelling [PITH_FULL_IMAGE:figures/full_fig_p012_1.png] view at source ↗
Figure 2
Figure 2. ROC for all three datasets In [PITH_FULL_IMAGE:figures/full_fig_p017_2.png] view at source ↗
Figure 3
Figure 3. LR classification results for the three datasets after pre-processing data at different [PITH_FULL_IMAGE:figures/full_fig_p019_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

64 extracted references · 59 canonical work pages

  1. [1]

    Https://www.govinfo.gov/content/pkg/CFR-2011-title29- vol4/xml/CFR-2011-title29-vol4-part1607.xml

    US Government, Uniform Guidelines on Employee Selection Procedures (1978), 2011. Https://www.govinfo.gov/content/pkg/CFR-2011-title29- vol4/xml/CFR-2011-title29-vol4-part1607.xml

    work page 1978

  2. [2]

    Barocas, A

    S. Barocas, A. D. Selbst, Big Data’s Disparate Impact, California Law Review 104 (2016) 671

    work page 2016

  3. [3]

    Calders, I

    T. Calders, I. Zliobaite, Why Unbiased Computational Processes can Lead to Discriminative Decision Procedures, in: Discrimination and Privacy in the Information Society, Springer, 2013, pp. 43–57

    work page 2013

  4. [4]

    B. T. Luong, S. Ruggieri, F. Turini, k-NN as an Implementation of Situa- tion Testing for Discrimination Discovery and Prevention, in: Proceedings of the 17th ACM SIGKDD International Conference on Knowledge Discov- ery and Data Mining, 2011, pp. 502–510

    work page 2011

  5. [5]

    Calmon, D

    F. Calmon, D. Wei, B. Vinzamuri, K. N. Ramamurthy, K. R. Varshney, Optimized Pre-Processing for Discrimination Prevention, in: Advances in Neural Information Processing Systems, 2017, pp. 3995–4004

    work page 2017

  6. [6]

    Feldman, S

    M. Feldman, S. A. Friedler, J. Moeller, C. Scheidegger, S. Venkatasubra- manian, Certifying and Removing Disparate Impact, in: Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, ACM, 2015, pp. 259–268. 27 Table D.8: Some discriminatory subgroups along with their discrimination value(discl), de- tected by KL...

    work page 2015

  7. [7]

    Madras, E

    D. Madras, E. Creager, T. Pitassi, R. Zemel, Learning Adversarially Fair and Transferable Representations, in: J. Dy, A. Krause (Eds.), Proceed- ings of the 35th International Conference on Machine Learning (MLR), volume 80 ofPMLR, PMLR, Stockholmsmässan, Stockholm Sweden, 2018, pp. 3384–3393. URL: http://proceedings.mlr.press/v80/madras18a. html

    work page 2018

  8. [8]

    E. Raff, J. Sylvester, S. Mills, Fair Forests: Regularized Tree Induction to Minimize Model Bias, in: Proceedings of the 2018 AAAI/ACM Conference on AI, Ethics, and Society, 2018, pp. 243–250

    work page 2018

Show all 64 references
  1. [9]

    Kamiran, T

    F. Kamiran, T. Calders, M. Pechenizkiy, Discrimination Aware Decision Tree Learning, in: Proceedings of the IEEE 10th International Conference on Data Mining (ICDM), IEEE, 2010, pp. 869–874

    2010

  2. [10]

    Aghaei, M

    S. Aghaei, M. J. Azizi, P. Vayanos, Learning Optimal and Fair Decision Tees for Non-Discriminative Decision-Making, in: Proceedings of the AAAI Conference on Artificial Intelligence, volume 33, 2019, pp. 1418–1426

    2019

  3. [11]

    Valdivia, J

    A. Valdivia, J. Sánchez-Monedero, J. Casillas, How Fair Can We Go in Machine Learning? Assessing the Boundaries of Accuracy and Fairness, International Journal of Intelligent Systems (2021) 1–25. URL:https:// doi.org/10.1002/int.22354

    2021 doi

  4. [12]

    P. K. Lohia, K. Natesan Ramamurthy, M. Bhide, D. Saha, K. R. Varshney, R.Puri, BiasMitigationPost-processingforIndividualandGroupFairness, in: Proceedings of the 2019 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP, 2019, pp. 2847–2851. doi:10....

    2019

  5. [13]

    Hardt, E

    M. Hardt, E. Price, N. Srebro, et al., Equality of Opportunity in Supervised Learning, in: Proceedings of the 30th International Conference on Neural Information Processing Systems, 2016, pp. 3323–3331

    2016

  6. [14]

    Awasthi, M

    P. Awasthi, M. Kleindessner, J. Morgenstern, Equalized Odds Postprocess- ing Under Imperfect Group Information, in: Proceedings of the Interna- tional Conference on Artificial Intelligence and Statistics, PMLR, 2020, pp. 1770–1780

    2020

  7. [15]

    Gutierrez, J.-Y

    P. Gutierrez, J.-Y. Gérardy, Causal Inference and Uplift Modelling: A Re- viewofthe Literature, in: Proceedingsofthe2016InternationalConference on Predictive Applications and APIs, PMLR, 2017, pp. 1–13

    2017

  8. [16]

    B.Becker, R.Kohavi, Adult, UCIMachineLearningRepository, 1996.DOI: https://doi.org/10.24432/C5XW20

    1996 doi

  9. [17]

    Larson, M

    J. Larson, M. Roswell, V. Atlidakis, COMPASS Recidivism Dataset, 2016. URL: https://github.com/propublica/compas-analysis. 29

    2016

  10. [18]

    D. Dua, C. Graff, German Credit Dataset, UCI Machine Learning Reposi- tory, 2017. URL:http://archive.ics.uci.edu/ml

    2017

  11. [19]

    Kamiran, T

    F. Kamiran, T. Calders, Classifying Without Discriminating, in: Pro- ceedings of the 2nd International Conference on Computer, Control and Communication, IC4, IEEE, 2009, pp. 1–6

    2009

  12. [20]

    Kamiran, T

    F. Kamiran, T. Calders, Data Preprocessing Techniques for Classification Without Discrimination, Knowledge and Information Systems 33 (2012) 1–33

    2012

  13. [21]

    M. Wan, D. Zha, N. Liu, N. Zou, In-processing Modeling Techniques for Machine Learning Fairness: A Survey, ACM Transactions on Knowledge Discovery from Data 17 (2023) 1–27

    2023

  14. [22]

    Jiang, A

    R. Jiang, A. Pacchiano, T. Stepleton, H. Jiang, S. Chiappa, Wasserstein Fair Classification, in: Uncertainty in artificial intelligence, PMLR, 2020, pp. 862–872

    2020

  15. [23]

    Agarwal, M

    A. Agarwal, M. Dudík, Z. S. Wu, Fair Regression: Quantitative Definitions and Reduction-based Algorithms, in: International Conference on Machine Learning, PMLR, 2019, pp. 120–129

    2019

  16. [24]

    Saxena, S

    S. Saxena, S. Jain, Exploring and Mitigating Gender Bias in Book Recom- mender Systems with Explicit Feedback, Journal of Intelligent Information Systems (2024). doi:10.1007/s10844-023-00827-8

    2024 doi

  17. [25]

    García-Soriano, F

    D. García-Soriano, F. Bonchi, Maxmin-fair Ranking: Individual Fair- ness under Group-fairness Constraints, in: Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining, 2021, pp. 436–446

    2021

  18. [26]

    Lahoti, A

    P. Lahoti, A. Beutel, J. Chen, K. Lee, F. Prost, N. Thain, X. Wang, E. Chi, Fairness Without Demographics through Adversarially Reweighted Learn- ing, Advances in neural information processing systems 33 (2020) 728–740

    2020

  19. [27]

    S. Garg, V. Perot, N. Limtiaco, A. Taly, E. H. Chi, A. Beutel, Counterfac- tual Fairness in Text Classification through Robustness, in: Proceedings of the 2019 AAAI/ACM Conference on AI, Ethics, and Society, 2019, pp. 219–226

    2019

  20. [28]

    Sweeney, M

    C. Sweeney, M. Najafian, Reducing Sentiment Polarity for Demographic Attributes in Word Embeddings using Adversarial Learning, in: Proceed- ings of the 2020 Conference on Fairness, Accountability, and Transparency, 2020, pp. 359–368

    2020

  21. [29]

    B. H. Zhang, B. Lemoine, M. Mitchell, Mitigating Unwanted Biases with Adversarial Learning, in: Proceedings of the 2018 AAAI/ACM Conference on AI, Ethics, and Society, 2018, pp. 335–340. 30

    2018

  22. [30]

    H. Kim, S. Shin, J. Jang, K. Song, W. Joo, W. Kang, I.-C. Moon, Coun- terfactual Fairness with Disentangled Causal Effect Variational Autoen- coder, in: Proceedings of the AAAI Conference on Artificial Intelligence, volume 35, 2021, pp. 8128–8136

    2021

  23. [31]

    S. Park, S. Hwang, D. Kim, H. Byun, Learning Disentangled Representa- tion for Fair Facial Attribute Classification via Fairness-aware Information Alignment, in: Proceedings of the AAAI Conference on Artificial Intelli- gence, volume 35, 2021, pp. 2403–2411

    2021

  24. [32]

    Cheng, W

    P. Cheng, W. Hao, S. Yuan, S. Si, L. Carin, FairFil: Contrastive Neural De- biasing Method for Pretrained Text Encoders, in: International Conference on Learning Representations, 2020, pp. 1–12

    2020

  25. [33]

    C. Zhou, J. Ma, J. Zhang, J. Zhou, H. Yang, Contrastive Learning for Debiased Candidate Generation in Large-scale Recommender Systems, in: Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discov- ery & Data Mining, 2021, pp. 3985–3995

    2021

  26. [34]

    Kamiran, A

    F. Kamiran, A. Karim, X. Zhang, Decision Theory for Discrimination- Aware Classification, in: Proceedings of the IEEE 12th International Con- ference on Data Mining, ICDM’12, IEEE, 2012, pp. 924–929. doi:10.1109/ ICDM.2012.45

    2012

  27. [35]

    Zhang, A

    W. Zhang, A. Bifet, FEAT: A Fairness-Enhancing and Concept-Adapting Decision Tree Classifier, in: Discovery Science, volume 12323 of Lec- ture Notes in Computer Science, Springer International Publishing, Cham, Switzerland, 2020, pp. 175–189

    2020

  28. [36]

    Castelnovo, A

    A. Castelnovo, A. Cosentini, L. Malandri, F. Mercorio, M. Mezzanzanica, FFTree: A Flexible Tree to Handle Multiple Fairness Criteria, Information Processing & Management 59 (2022) 103099

    2022

  29. [37]

    Jaskowski, S

    M. Jaskowski, S. Jaroszewicz, Uplift Modeling for Clinical Trial Data, in: Proceedings of the 29th International Coference on International Confer- ence on Machine Learning (ICML), Workshop on Clinical Data Analysis, Omnipress, Madison, WI, United States, 2012, pp. 1–8

    2012

  30. [38]

    J. Gao, X. Zheng, D. Wang, Z. Huang, B. Zheng, K. Yang, UTBoost: A Tree-boosting based System for Uplift Modeling, CoRR abs/2312.02573 (2023). doi:10.48550/ARXIV.2312.02573. arXiv:2312.02573

    2023 doi arXiv

  31. [39]

    Hansotia, B

    B. Hansotia, B. Rukstales, Incremental Value Modeling, Journal of Inter- active Marketing 16 (2002) 1–35

    2002

  32. [40]

    Su, C.-L

    X. Su, C.-L. Tsai, H. Wang, D. M. Nickerson, B. Li, Subgroup Analysis via Recursive Partitioning, Journal of Machine Learning Research 10 (2009) 141–158. 31

    2009

  33. [41]

    Rzepakowski, S

    P. Rzepakowski, S. Jaroszewicz, Decision Trees for Uplift Modeling, in: Proceedings of the 10th International Conference on Data Mining, ICDM’10, IEEE, 2010, pp. 441–450

    2010

  34. [42]

    Y. He, K. Burghardt, S. Guo, K. Lerman, Inherent Trade-offs in the Fair Allocation of Treatments, 2020.arXiv:2010.16409

    2020 arXiv

  35. [43]

    Athey, G

    S. Athey, G. Imbens, Recursive Partitioning for Heterogeneous Causal Effects, Proceedings of the National Academy of Sciences 113 (2016) 7353–

    2016

  36. [44]

    N. Jo, S. Aghaei, J. Benson, A. Gomez, P. Vayanos, Learning Optimal Fair Decision Trees: Trade-offs Between Interpretability, Fairness, and Ac- curacy, in: Proceedings of the 2023 AAAI/ACM Conference on AI, Ethics, and Society, AIES ’23, Association for Computing Machinery, Ne...

    2023

  37. [45]

    Rudin, C

    C. Rudin, C. Chen, Z. Chen, H. Huang, L. Semenova, C. Zhong, In- terpretable machine learning: Fundamental principles and 10 grand chal- lenges, Statistics Surveys 16 (2022) 1 – 85. URL:https://doi.org/10. 1214/21-SS133. doi:10.1214/21-SS133

    2022 doi

  38. [46]

    Scarpato, A

    N. Scarpato, A. Nourbakhsh, P. Ferroni, S. Riondino, M. Roselli, F. Fallucchi, P. Barbanti, F. Guadagni, F. M. Zanzotto, Evaluat- ing Explainable Machine Learning Models for Clinicians, Cognitive Computation 16 (2024) 1436–1446. URL: https://doi.org/10.1007/ s12559-024-10297-x...

    2024 doi

  39. [47]

    Molnar, Interpretable Machine Learning, 2 ed., 2022

    C. Molnar, Interpretable Machine Learning, 2 ed., 2022. URL:https:// christophm.github.io/interpretable-ml-book

    2022

  40. [48]

    D. V. Carvalho, E. M. Pereira, J. S. Cardoso, Machine learning inter- pretability: A survey on methods and metrics, Electronics 8 (2019) 832

    2019

  41. [49]

    Csiszár, P

    I. Csiszár, P. C. Shields, et al., Information Theory and Statistics: A Tutorial, Foundations and Trends in Communications and Information Theory 1 (2004) 417–528

    2004

  42. [50]

    Sołtys, S

    M. Sołtys, S. Jaroszewicz, P. Rzepakowski, Ensemble Methods for Uplift Modeling, Data Mining and Knowledge Discovery 29 (2015) 1531–1559

    2015

  43. [51]

    T. S. Han, K. Kobayashi, Mathematics of Information and Coding (Trans- lations of Mathematical Monographs), American Mathematical Society, USA, 2001

    2001

  44. [52]

    S. Yu, A. Shaker, F. Alesiani, J. Principe, Measuring the Discrepancy be- tween Conditional Distributions: Methods, Properties and Applications, in: C. Bessiere (Ed.), Proceedings of the 29th International Joint Confer- ence on Artificial Intelligence, IJCAI’20, IJCAI Organiza...

    2020 doi

  45. [53]

    Kullback, Letter to the Editor: The Kullback-Leibler Distance, The American Statistician (1987) 338–341

    S. Kullback, Letter to the Editor: The Kullback-Leibler Distance, The American Statistician (1987) 338–341

    1987

  46. [54]

    Hajian, J

    S. Hajian, J. Domingo-Ferrer, A Methodology for Direct and Indirect Dis- crimination Prevention in Data Mining, IEEE Transactions on Knowledge and Data Engineering 25 (2013) 1445–1459

    2013

  47. [55]

    Ruggieri, Using t-Closeness Anonymity to Control for Non- Discrimination, Transactions on Data Privacy 7 (2014) 99–129

    S. Ruggieri, Using t-Closeness Anonymity to Control for Non- Discrimination, Transactions on Data Privacy 7 (2014) 99–129

    2014

  48. [56]

    R. K. E. Bellamy, K. Dey, M. Hind, S. C. Hoffman, S. Houde, K. Kannan, P.Lohia, J.Martino, S.Mehta, A.Mojsilović, S.Nagar, K.N.Ramamurthy, J. Richards, D. Saha, P. Sattigeri, M. Singh, K. R. Varshney, Y. Zhang, AI Fairness 360: An Extensible Toolkit for Detecting, Understandin...

    2019

  49. [57]

    J. M. Klusowski, P. M. Tian, Large scale prediction with decision trees, Journal of the American Statistical Association 119 (2023) 525–

    2023

  50. [58]

    Equal Employment Opportunity Commission637, Discrimination by Type, 2024

    U.S. Equal Employment Opportunity Commission637, Discrimination by Type, 2024. Https://www.eeoc.gov/discrimination-type

    2024

  51. [59]

    Falk, Inequalities of JW Gibbs, American Journal of Physics 38 (1970) 858–869

    H. Falk, Inequalities of JW Gibbs, American Journal of Physics 38 (1970) 858–869

    1970

  52. [60]

    J. R. Quinlan, Induction of Decision Trees, Machine learning 1 (1986) 81–106

    1986

  53. [61]

    Breiman, J

    L. Breiman, J. Friedman, C. Stone, R. Olshen, Classification and Regres- sion Trees, Taylor & Francis, Florida, USA, 1984. URL:https://books. google.com.pk/books?id=JwQx-WOmSyQC

    1984

  54. [62]

    Pedregosa, G

    F. Pedregosa, G. Varoquaux, A. Gramfort, V. Michel, B. Thirion, O. Grisel, M. Blondel, P. Prettenhofer, R. Weiss, V. Dubourg, J. Vanderplas, A. Pas- sos, D. Cournapeau, M. Brucher, M. Perrot, E. Duchesnay, Scikit-learn: Machine learning in Python, Journal of Machine Learning R...

    2011

  55. [537]

    URL: https://doi.org/10.1080/01621459.2022.2126782. doi:10. 1080/01621459.2022.2126782

    2022

  56. [7360]

    doi:10.1073/pnas.1510489113

    doi

Pith tools

Reviewed August 9, 2026 · model on record in the stance chip above.