Pith. sign in

REVIEW 2 major objections 6 minor 29 references

Privilege Scores

T0 review · 2 major / 6 minor · reviewed 2026-08-09 · deepseek-v4-flash

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

Pith's one-line read This paper proposes a privilege score: the gap between a person's predicted outcome in the real world and in a fair world in which the protected attribute has no causal effect, decomposed into per-path contributions.

desk verdict A clean formalization of privilege as a prediction gap, with a useful Shapley-style decomposition; the path-level attributions rest on assumed DAGs that even the paper's own misspecification simulation shows are fragile, so the real-world policy conclusions overshoot. read the letter →

arxiv 2502.01211 v1 pith:Z7KXDBM6 submitted 2025-02-03 cs.LG stat.ML

Ludwig Bothmann , Philip A. Boustani , Jose M. Alvarez , Giuseppe Casalicchio , Bernd Bischl , Susanne Dandl This is my paper · ORCID
classification cs.LGstat.ML
keywords privilegescoresfairness-awaremachinelearningbias-transformingfairnessFiNDworldShapleyvaluescausalmediationcounterfactualreasoningmodelinterpretability
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 privilege scores, a way to put a number on how much a protected attribute such as race or gender helps or hurts an individual in an automated decision. For each person, the score is the difference between their predicted probability of the positive outcome in the real world and their predicted probability in a 'fair world' in which the protected attribute has no causal effect on the outcome. The authors argue this gives bias-transforming fairness an explicit target: it identifies individuals who qualify for affirmative action and quantifies, at the group level, how much privilege exists and through which features it flows. They also show how to decompose the score into contributions of specific causal paths and provide confidence intervals, then demonstrate the approach on mortgage lending and law school admission data.

What carries the argument

The FiND world is the fair-world baseline: a fictional world where protected attributes have no causal effect on the target, approximated by warping the descendants of the protected attribute to counterfactual values, here via fairadapt (quantile-preserving causal adaptation) and residual-based warping. Privilege score contributions are then computed as Shapley values, with the value function v(S) = π̂(x) − π̂(x_S) defined over coalitions of the arrows starting at the protected attribute, so each arrow receives its average marginal contribution to the total privilege score; an efficiency theorem guarantees the contributions sum to the warping part of the score, while the intercepts absorb the remaining difference between the real-world and warped-world models.

What would settle it

Take the paper's mortgage or law school data, compute privilege scores under two defensible causal graphs that differ in whether a plausible confounder is affected by the protected attribute, and count how many individuals cross an affirmative-action threshold under each graph; if the set of flagged individuals changes materially, the score's practical conclusions hinge on the unvalidated DAG assumption.

Watch

Extended reading notes

Core claim

The central claim is that privilege is measurable as δ(x, x^F) = π(x) − ψ(x^F), the difference between the probability a model assigns to an individual in the real world and the probability assigned in the normatively desired fair world (FiND world) in which the protected attribute has no direct or indirect effect on the target. The paper argues this individual-level score answers what bias-transforming methods are actually rectifying, and that a Shapley-style decomposition, the privilege score contribution (PSC), attributes the score to each PA-to-feature causal arrow plus a global and individual intercept that captures direct effects and unobserved mediators. It provides bootstrap confidence intervals for both scores, and its simulations show low bias for two warping methods while its real-world analyses find, for example, that racial privilege in law school admission operates mainly through LSAT scores rather than through unobserved factors.

Load-bearing premise

The causal graph describing how the protected attribute influences other features must be correct, because the whole fair world and therefore every privilege score is built on it.

Editorial extensions

If this is right

  • An individual with a PS whose confidence interval excludes zero has statistically detectable privilege or disadvantage in the decision, which can be used to substantiate a claim of discrimination or to qualify for affirmative action.
  • Averaging PS over a group quantifies group-level privilege, and regressing PS on real-world features yields location- and context-specific estimates of racial and gender bias, as in the mortgage analysis.
  • PSC importances decompose privilege into mediators, so a policy aimed at equalizing outcomes can target the dominant path, such as LSAT scores in the law school analysis.
  • Because the score is defined as a difference of two probability predictions, the same framework applies at the dataset, model, or decision stage and can be extended to regression outcomes.

Reading between the lines

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

  • One testable extension, not explored in the paper, is to compare PS across two plausible DAGs on the same data; if individual rankings and policy conclusions are stable, practitioners could trust the score without knowing the true causal graph.
  • The intercept terms could be read as a diagnostic for unmeasured mediators or direct discrimination, but that reading inherits the warping method's assumptions and should be validated by collecting additional features.
  • The score's definition is agnostic to how the fair world is approximated, so non-causal warping, optimal transport, or generative counterfactual models could be plugged in as estimation methods.
  • Current warping methods handle one protected attribute at a time and cannot fully perform partial warpings for features with multiple PA-induced paths, which limits the score's use for intersectional fairness until those methods advance.
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

2 major / 6 minor

Summary. The paper introduces the privilege score (PS), δ = π(x) − ψ(xF), which compares an individual's predicted probability of a positive outcome under the real-world model with the prediction under a counterfactual 'fair world' (FiND world) where the protected attribute has no causal effect. Estimation proceeds by warping the feature distribution (fairadapt or residual-based warping) to approximate the FiND world, learning models in both worlds, and taking the difference. The paper proposes privilege score contributions (PSCs), a Shapley-value decomposition of PS that assigns the privilege to individual PA-to-feature arrows and an intercept term, and provides bootstrap confidence intervals. The method is evaluated in a simulation study (including a DAG misspecification scenario) and on HMDA mortgage data and law school data, yielding conclusions such as a strong racial privilege mediated by loan purpose in Louisiana and by LSAT in law school admissions.

Significance. If the counterfactual FiND baseline can be credibly constructed, PS provides a principled and interpretable tool for bias-transforming fairness: it operationalizes the 'status quo' as a measurable individual-level quantity, and the Shapley-based PSCs offer a transparent way to trace privilege through mediators. The paper includes a machine-checked efficiency proof, bootstrap CIs, and a simulation study with a misspecification arm, which are notable strengths. The major weakness is that the real-world conclusions depend on assumed causal DAGs; the paper's own misspecification results show that path-level estimates are sensitive to even a single omitted edge, and no sensitivity analysis is provided for the real-world applications. Thus the methodological framework is valuable, but the path-attribution conclusions and policy recommendations require additional robustness work.

major comments (2)
  1. [5.2 (Tables 4-5) and Appendix B.2 (Table 8)] The real-world path-level conclusions lack the robustness evidence that the paper's own simulation suggests is necessary. In the misspecification scenario SM (Table 8), a single omitted edge A→C yields a bias of -0.04 for the global intercept δg with coverage falling to 0.31 (res-based warping) and a bias of -0.018 for the path contribution γ1(x); fairadapt shows a bias of 0.037 for δ. The law-school analysis (Section 5.2, Table 5) nevertheless asserts that Black students' negative privilege is mainly mediated by LSAT (γ2(x) = -0.111) and recommends equalizing LSAT scores, based on the assumed DAG in Figure 3b without any sensitivity analysis. Since the central claim is that PSCs identify the causal paths through which privilege operates, the authors should either add a sensitivity analysis over plausible alternative DAGs (e.g., a direct A→Y edge, unmeasured confounders affecting both LSAT and bar passage) or explicitly bound how much the headline γj estimates could change, and temper the policy claims accordingly.
  2. [5.2 (Law school data)] The interpretation of the PSC as a policy recommendation is an overreach. The PSC γ2(x) measures the change in the real-world model's prediction when LSAT is warped from its factual to its fair-world value, holding the rest of the pipeline fixed; it does not evaluate the intervention 'equalize LSAT scores', which would alter the joint distribution of features and outcomes beyond the model's input. The statement that 'policies aimed at effectively increasing racial equality should aim at equalizing LSAT scores' should be presented as a hypothesis for policy analysis rather than a direct conclusion of the PS decomposition, unless additional evidence is provided.
minor comments (6)
  1. [5.1, Table 2] The coverage values of 0.969-0.977 in the correctly specified scenario are well above the nominal 0.9 and are described as 'slightly above'; this overcoverage deserves a sentence of explanation (e.g., due to the bootstrap or the conservative random forest predictions).
  2. [5.1, Table 3] The coverage of γ2(x) for res-based warping is 0.866, below the nominal 0.9; the text calls this 'slightly too low' — please quantify the Monte Carlo uncertainty of the coverage estimate and discuss the miscalibration.
  3. [4, A.2] The paper claims to show efficiency (Theorem 4.1) and other axioms, but symmetry and dummy are only asserted in one sentence; a concise explicit proof would make the axiomatic claim self-contained.
  4. [2 and Appendix B] The residual-based warping method is a key estimation tool, but the main text gives no algorithmic description and refers only to Bothmann et al. (2023); a short pseudo-code or a precise definition in the appendix would improve reproducibility.
  5. [5.2, first paragraph] 'Data are splitted into 80%/20% train/test sets' should read 'split'; also, the use of 'CIs' for confidence intervals is inconsistent in places (sometimes 'CI's').
  6. [Table 4 and surrounding text] The paper notes that formal tests for PSC importance are not yet developed, but earlier the abstract and Section 1 advertise that it 'provides confidence intervals for both PS and PSCs'; clarify that the CIs apply to individual-level quantities and not to the aggregated importance values.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the privilege-score definition, its warping-based estimation, and the Shapley decomposition are self-contained, with prior-work citations serving as non-load-bearing building blocks.

full rationale

The paper's central derivation is not circular. Definition 3.1 defines the privilege score as δ = π(x) − ψ(xF), and the PSC construction (Eq. 3-4) is an explicit Shapley decomposition of the estimated difference δ̂ = π̂(x) − φ̂(x̃), with efficiency proven in Theorem 4.1. The decomposition is a mathematical identity relative to the chosen value function v(S) = π̂(x) − π̂(xS); it does not smuggle the conclusion into the input. The warping methods (fairadapt and res-based) are prior building blocks, cited from Plečko & Meinshausen (2020) and Bothmann et al. (2023), and the paper tests them in a simulation study against known ground truth, which provides independent validation of the estimation pipeline. The FiND-world concept is cited from Bothmann et al. (2024), but the paper explicitly states the PS framework is agnostic to the precise definition of the fair world, so this citation is not load-bearing for the technical derivation. The acknowledged limitation of DAG misspecification (Section 5, Discussion) is a validity/robustness concern, not a circularity: the paper's own misspecification simulation (SM) quantifies the degradation, and the real-world DAGs are stated assumptions rather than conclusions derived from the method. No fitted parameter is relabeled as a prediction, and no 'uniqueness' theorem is invoked to force the choice of fair world. The self-citations are thus normal prior-work references, and the central claim has independent content.

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

The central claim rests on the FiND world construct (from prior work), the accuracy of warping methods, the correctness of assumed DAGs, and standard Shapley axioms. The paper does not introduce new physical or conceptual entities beyond the privilege score quantity itself, which is a function of existing models. The main burden is the normative and causal assumptions, not hidden free parameters.

assumptions (5)
  • domain assumption The FiND world exists and is normatively desired; the protected attribute has no direct or indirect causal effect on the target.
    Section 2: 'we follow the philosophical rationale of Bothmann et al. (2024) who propose a fictitious, normatively desired (FiND) world, where the PAs have no direct nor indirect causal effect on the target.' This is a normative assumption imported from self-cited work.
  • domain assumption The warping methods (fairadapt and residual-based) produce a valid approximation of the FiND world.
    Section 2 states that both methods are causal and 'very closely adopt the philosophy of the FiND world concept'; the simulation only tests one misspecification scenario, so validity is assumed.
  • domain assumption The causal DAGs in Figures 2 and 3 correctly represent the true causal structure for the respective applications.
    Section 5.2: 'Figure 3a shows assumed DAGs.' The paper acknowledges sensitivity to misspecification but real-world results rely on these assumptions.
  • standard math Shapley axioms (efficiency, symmetry, dummy) provide a fair allocation of privilege contributions.
    Section 4 frames PSCs via Shapley values; these are standard cooperative game theory axioms, and the efficiency proof in Appendix A.2 is correct.
  • domain assumption The ML models π̂ and φ̂ are unbiased estimators of π and φ, and the warped world equals the FiND world.
    Section 3 theoretical analysis: 'With unbiased ML models and with a perfect warping method, δ̂ is an unbiased estimator for δ.' This is a stated conditional assumption, not proven.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Privilege Scores." pith.science (2026). https://pith.science/paper/Z7KXDBM6

@misc{pith2026250201211,
  author       = {Pith},
  title        = {Pith review of: Privilege Scores},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/Z7KXDBM6}},
  note         = {Machine review of arXiv:2502.01211}
}
read the original abstract

Bias-transforming methods of fairness-aware machine learning aim to correct a non-neutral status quo with respect to a protected attribute (PA). Current methods, however, lack an explicit formulation of what drives non-neutrality. We introduce privilege scores (PS) to measure PA-related privilege by comparing the model predictions in the real world with those in a fair world in which the influence of the PA is removed. At the individual level, PS can identify individuals who qualify for affirmative action; at the global level, PS can inform bias-transforming policies. After presenting estimation methods for PS, we propose privilege score contributions (PSCs), an interpretation method that attributes the origin of privilege to mediating features and direct effects. We provide confidence intervals for both PS and PSCs. Experiments on simulated and real-world data demonstrate the broad applicability of our methods and provide novel insights into gender and racial privilege in mortgage and college admissions applications.

Figures

Figures reproduced from arXiv: 2502.01211 by the authors.

Figure 1
Figure 1. PS and PS contributions for Amina. Rectifying a non-neutral status quo is increasingly stud￾ied in fairML (e.g., Alvarez et al., 2024; Mittelstadt et al., 2024; Russo et al., 2024). Motivated by discussions on non￾discrimination law (e.g., Kohler-Hausmann, 2019; Wachter et al., 2021; Weerts et al., 2023), including affirmative action (Romei & Ruggieri, 2014), a growing body of fairML work targets substantive equalit… view at source ↗
Figure 2
Figure 2. Exemplary DAGs. Real world: all arrows exist. [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. DAGs for (a) mortgage and (b) law school data. [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (10 more)
Figure 4
Figure 4. Figure 4: LA: PS and PSCs for a non-white female person, [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: Exemplary DAG for partial warping. In a DAG such as in [PITH_FULL_IMAGE:figures/full_fig_p014_5.png]
Figure 6
Figure 6. Figure 6: LA – res-based – Individuals with smallest (left) and largest (right) PS. [PITH_FULL_IMAGE:figures/full_fig_p021_6.png]
Figure 7
Figure 7. Figure 7: LA – fairadapt – Individuals with smallest (left) and largest (right) PS. [PITH_FULL_IMAGE:figures/full_fig_p022_7.png]
Figure 8
Figure 8. Figure 8: WI – res-based – Individuals with smallest (left) and largest (right) PS. [PITH_FULL_IMAGE:figures/full_fig_p026_8.png]
Figure 9
Figure 9. Figure 9: WI – fairadapt – Individuals with smallest (left) and largest (right) PS. [PITH_FULL_IMAGE:figures/full_fig_p027_9.png]
Figure 10
Figure 10. Figure 10: NY – res-based – Individuals with smallest (left) and largest (right) PS. [PITH_FULL_IMAGE:figures/full_fig_p031_10.png]
Figure 11
Figure 11. Figure 11: NY – fairadapt – Individuals with smallest (left) and largest (right) PS. [PITH_FULL_IMAGE:figures/full_fig_p032_11.png]
Figure 12
Figure 12. Figure 12: PS Contribution for the Lawschool data with [PITH_FULL_IMAGE:figures/full_fig_p036_12.png]
Figure 13
Figure 13. Figure 13: PS Contribution for the Lawschool data with [PITH_FULL_IMAGE:figures/full_fig_p037_13.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

29 extracted references · 14 canonical work pages

  1. [1]

    Alvarez, J. M. and Ruggieri, S. Counterfactual Situation Testing : Uncovering Discrimination under Fairness given the Difference . In Proceedings of the 3rd ACM Conference on Equity and Access in Algorithms , Mechanisms , and Optimization , EAAMO '23, pp.\ 1--11, New York, NY, USA, October 2023. Association for Computing Machinery. doi:10.1145/3617694.362...

    arXiv 2023

  2. [2]

    M., Bringas-Colmenarejo, A., Elobaid, A., Fabbrizzi, S., Fahimi, M., Ferrara, A., Ghodsi, S., Mougan, C., Papageorgiou, I., Reyero, P., et al

    Alvarez, J. M., Bringas-Colmenarejo, A., Elobaid, A., Fabbrizzi, S., Fahimi, M., Ferrara, A., Ghodsi, S., Mougan, C., Papageorgiou, I., Reyero, P., et al. Policy advice and best practices on bias and fairness in AI . Ethics and Information Technology, 26 0 (2): 0 31, 2024. doi:10.1007/s10676-024-09746-w. URL https://link.springer.com/article/10.1007/s1067...

    doi:10.1007/s10676-024-09746-w 2024

  3. [4]

    Causal Fair Machine Learning via Rank - Preserving Interventional Distributions

    Bothmann, L., Dandl, S., and Schomaker, M. Causal Fair Machine Learning via Rank - Preserving Interventional Distributions . In Proceedings of the 1st Workshop on Fairness and Bias in AI co-located with 26th European Conference on Artificial Intelligence ( ECAI 2023) . CEUR Workshop Proceedings, October 2023. URL https://ceur-ws.org/Vol-3523/

    work page 2023

  4. [5]

    What Is Fairness ? On the Role of Protected Attributes and Fictitious Worlds

    Bothmann, L., Peters, K., and Bischl, B. What Is Fairness ? On the Role of Protected Attributes and Fictitious Worlds . 2024. doi:10.48550/arXiv.2205.09622. URL http://arxiv.org/abs/2205.09622

    arXiv doi:10.48550/arxiv.2205.09622 2024

  5. [6]

    Fair Prediction with Disparate Impact : A Study of Bias in Recidivism Prediction Instruments

    Chouldechova, A. Fair Prediction with Disparate Impact : A Study of Bias in Recidivism Prediction Instruments . Big Data, 5 0 (2): 0 153--163, 2017. doi:10.1089/big.2016.0047. URL https://www.liebertpub.com/doi/10.1089/big.2016.0047

    arXiv 2017

  6. [7]

    Fairness is not static: deeper understanding of long term fairness via simulation studies

    D'Amour, A., Srinivasan, H., Atwood, J., Baljekar, P., Sculley, D., and Halpern, Y. Fairness is not static: deeper understanding of long term fairness via simulation studies. In Proceedings of the 2020 Conference on Fairness , Accountability , and Transparency , FAT * '20, pp.\ 525--534. ACM , 2020. doi:10.1145/3351095.3372878. URL https://doi.org/10.1145...

    arXiv 2020

  7. [8]

    K., Bothmann, L., Wright, M

    Ewald, F. K., Bothmann, L., Wright, M. N., Bischl, B., Casalicchio, G., and K \"o nig, G. A guide to feature importance methods for scientific inference. In Longo, L., Lapuschkin, S., and Seifert, C. (eds.), Explainable Artificial Intelligence, pp.\ 440--464, Cham, 2024. Springer Nature Switzerland. doi:10.1007/978-3-031-63797-1_22. URL https://link.sprin...

    doi:10.1007/978-3-031-63797-1_22 2024

  8. [9]

    Friedman, J. H. and Fisher, N. I. Bump hunting in high-dimensional data. Statistics and Computing, 9 0 (2): 0 123--143, 1999. doi:10.1023/A:1008894516817. URL https://link.springer.com/article/10.1023/A:1008894516817

    work page doi:10.1023/a:1008894516817 1999

Show all 29 references
  1. [10]

    and Chen, Y

    Hu, L. and Chen, Y. A short-term intervention for long-term fairness in the labor market. In Proceedings of the 2018 World Wide Web Conference, pp.\ 1389--1398. ACM , 2018. doi:10.1145/3178876.3186044. URL https://doi.org/10.1145/3178876.3186044

    2018

  2. [11]

    Inherent Trade - Offs in the Fair Determination of Risk Scores

    Kleinberg, J., Mullainathan, S., and Raghavan, M. Inherent Trade - Offs in the Fair Determination of Risk Scores . In Papadimitriou, C. H. (ed.), 8th Innovations in Theoretical Computer Science Conference ( ITCS 2017) , volume 67 of Leibniz International Proceedings in Informa...

    2017 doi

  3. [12]

    Eddie Murphy and the Dangers of Counterfactual Causal Thinking About Detecting Racial Discrimination

    Kohler-Hausmann, I. Eddie Murphy and the Dangers of Counterfactual Causal Thinking About Detecting Racial Discrimination . Northwestern University Law Review, 113 0 (5): 0 1163--1228, 2019. URL https://scholarlycommons.law.northwestern.edu/nulr/vol113/iss5/6

    2019

  4. [13]

    J., Loftus, J., Russell, C., and Silva, R

    Kusner, M. J., Loftus, J., Russell, C., and Silva, R. Counterfactual Fairness . In Guyon, I., Luxburg, U. V., Bengio, S., Wallach, H., Fergus, R., Vishwanathan, S., and Garnett, R. (eds.), Advances in Neural Information Processing Systems , volume 30. Curran Associates, Inc., ...

    2017

  5. [14]

    Overcoming F airness T rade-offs via P re-processing: A C ausal P erspective

    Leininger, C., Rittel, S., and Bothmann, L. Overcoming F airness T rade-offs via P re-processing: A C ausal P erspective. 2025. doi:10.48550/arXiv.2501.14710. URL https://arxiv.org/abs/2501.14710

    2025 doi arXiv

  6. [15]

    M., Erion, G., Chen, H., DeGrave, A., Prutkin, J

    Lundberg, S. M., Erion, G., Chen, H., DeGrave, A., Prutkin, J. M., Nair, B., Katz, R., Himmelfarb, J., Bansal, N., and Lee, S.-I. From local explanations to global understanding with explainable AI for trees. Nature Machine Intelligence, 2 0 (1): 0 56--67, 2020. doi:10.1038/s4...

    2020 doi

  7. [16]

    Does enforcing fairness mitigate biases caused by subpopulation shift? In Ranzato, M., Beygelzimer, A., Dauphin, Y., Liang, P., and Vaughan, J

    Maity, S., Mukherjee, D., Yurochkin, M., and Sun, Y. Does enforcing fairness mitigate biases caused by subpopulation shift? In Ranzato, M., Beygelzimer, A., Dauphin, Y., Liang, P., and Vaughan, J. W. (eds.), Advances in Neural Information Processing Systems, volume 34, pp.\ 25...

    2021

  8. [17]

    The unfairness of fair machine learning: Leveling down and strict egalitarianism by default

    Mittelstadt, B., Wachter, S., and Russell, C. The unfairness of fair machine learning: Leveling down and strict egalitarianism by default. Michigan Technology Law Review, 30 0 (1): 0 3, 2024. URL https://ssrn.com/abstract=4331652

    2024

  9. [18]

    Causality: Models, Reasoning, and Inference

    Pearl, J. Causality: Models, Reasoning, and Inference. Cambridge University Press, 2nd edition, 2009

    2009

  10. [19]

    and Meinshausen, N

    Plečko, D. and Meinshausen, N. Fair Data Adaptation with Quantile Preservation . Journal of Machine Learning Research, 21: 0 1--44, 2020. URL http://jmlr.org/papers/v21/19-966.html

    2020

  11. [20]

    T., Lamba, H., and Ghani, R

    Rodolfa, K. T., Lamba, H., and Ghani, R. Empirical observation of negligible fairness-accuracy trade-offs in machine learning for public policy. Nature Machine Intelligence, 3 0 (10): 0 896--904, 2021. doi:10.1038/s42256-021-00396-x. URL https://www.nature.com/articles/s42256-...

    2021 doi

  12. [21]

    and Ruggieri, S

    Romei, A. and Ruggieri, S. A multidisciplinary survey on discrimination analysis. The Knowledge Engineering Review, 29 0 (5): 0 582--638, 2014. doi:10.1017/S0269888913000039. URL https://www.cambridge.org/core/product/D69E925AC96CDEC643C18A07F2A326D7

    2014 doi

  13. [22]

    M., Xu, W., Nguyen, D

    Russo, M., Jorgensen, M., Scott, K. M., Xu, W., Nguyen, D. H., Finocchiaro, J., and Olckers, M. Bridging research and practice through conversation: Reflecting on our experience. In Proceedings of the 4th ACM Conference on Equity and Access in Algorithms, Mechanisms, and Optim...

    2024

  14. [23]

    and Remmers, P

    Schw \" o bel, P. and Remmers, P. The long arc of fairness: Formalisations and ethical discourse. In Proceedings of the 2022 ACM Conference on Fairness, Accountability, and Transparency, FAccT '22, pp.\ 2179--2188. ACM , 2022. doi:10.1145/3531146.3534635. URL https://doi.org/1...

    2022

  15. [24]

    Shapley, L. S. A Value for n- Person Games . In Kuhn, H. W. and Tucker, A. W. (eds.), Contributions to the Theory of Games , Volume II , pp.\ 307--318. Princeton University Press, Princeton, 1953. doi:10.1515/9781400881970-018. URL https://doi.org/10.1515/9781400881970-018

    1953 doi

  16. [25]

    and Kononenko, I

    S trumbelj, E. and Kononenko, I. Explaining prediction models and individual predictions with feature contributions. Knowledge and Information Systems, 41 0 (3): 0 647--665, 2014. doi:10.1007/s10115-013-0679-x. URL https://link.springer.com/article/10.1007/s10115-013-0679-x

    2014 doi

  17. [26]

    Bias Preservation in Machine Learning : The Legality of Fairness Metrics Under EU Non - Discrimination Law

    Wachter, S., Mittelstadt, B., and Russell, C. Bias Preservation in Machine Learning : The Legality of Fairness Metrics Under EU Non - Discrimination Law . West Virginia Law Review, 123 0 (3): 0 735--790, 2021. doi:10.2139/ssrn.3792772. URL https://papers.ssrn.com/abstract=3792772

    2021 doi

  18. [27]

    P., and Pechenizkiy, M

    Weerts, H., Xenidis, R., Tarissan, F., Olsen, H. P., and Pechenizkiy, M. Algorithmic Unfairness through the Lens of EU Non - Discrimination Law : Or Why the Law is not a Decision Tree . In Proceedings of the 2023 ACM Conference on Fairness , Accountability , and Transparency ,...

    2023

  19. [28]

    Unlocking fairness: a trade-off revisited

    Wick, M., panda, s., and Tristan, J.-B. Unlocking fairness: a trade-off revisited. In Wallach, H., Larochelle, H., Beygelzimer, A., d Alch\' e -Buc, F., Fox, E., and Garnett, R. (eds.), Advances in Neural Information Processing Systems, volume 32. Curran Associates, Inc., 2019...

    2019

  20. [29]

    Wightman, L. F. LSAC National Longitudinal Bar Passage Study. LSAC Research Report Series ., 1998. URL https://archive.lawschooltransparency.com/reform/projects/investigations/2015/documents/NLBPS.pdf

    1998

  21. [30]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

Pith tools

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