pith. sign in

arxiv: 2605.22952 · v1 · pith:WD4ZHPM3new · submitted 2026-05-21 · 💻 cs.DB

Measuring Database Unfairness via Dependency Quantification Under Differential Privacy

Mariia Vologdin , Yuchao Tao , Amir Gilad This is my paper

Pith reviewed 2026-05-25 05:26 UTC · model grok-4.3

classification 💻 cs.DB
keywords differential privacydatabase fairnessunfairness quantificationmutual informationdata repairMaxSATtop-k contribution
0
0 comments X

The pith

A formal framework quantifies database unfairness under differential privacy via three measures meeting positivity, monotonicity, and DP computability.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper develops a framework to quantify unfairness in databases released under differential privacy, where added noise and restricted access hinder standard fairness checks. It identifies three core requirements any suitable measure must satisfy and instantiates them as a mutual information measure using a total variation proxy, a data-repair measure reduced to weighted MaxSAT, and a top-k tuple contribution measure. Privacy-preserving algorithms are supplied and their sensitivity, accuracy, and efficiency are analyzed. Experiments on real datasets show the measures approximate non-private versions and reveal bias under privacy constraints.

Core claim

We propose a formal framework for quantifying data unfairness under DP instantiated through three complementary measures that satisfy positivity, monotonicity, and DP computability, with privacy-preserving algorithms whose sensitivity, accuracy, and efficiency are analyzed.

What carries the argument

Three complementary measures (mutual information with total variation proxy, data repair via weighted MaxSAT, top-k tuple contribution) that each satisfy the three desiderata of positivity, monotonicity, and DP computability.

If this is right

  • The measures satisfy positivity, monotonicity, and DP computability while supporting privacy-preserving computation.
  • The measures approximate their non-private counterparts on multiple real-world datasets.
  • The measures quantify bias under privacy constraints and yield insights for data management.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • These measures could support routine fairness audits of government or corporate datasets released under DP.
  • The weighted MaxSAT reduction opens a route to applying combinatorial optimization inside privacy mechanisms for other fairness tasks.
  • A composite score that merges the three measures might prove more stable than any one alone when noise levels vary.

Load-bearing premise

The three desiderata of positivity, monotonicity, and DP computability are sufficient to define and capture the relevant notion of database unfairness under differential privacy.

What would settle it

If experiments on datasets with known injected biases show that the private measures deviate substantially from non-private unfairness scores beyond the analyzed accuracy bounds, the claim that the measures faithfully quantify unfairness under DP would be refuted.

Figures

Figures reproduced from arXiv: 2605.22952 by Amir Gilad, Mariia Vologdin, Yuchao Tao.

Figure 1
Figure 1. Figure 1: Values of the unfairness measures (log scale) on the [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Fairness values (Demographic Parity and Condi [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Behavior of unfairness measures as the Demo [PITH_FULL_IMAGE:figures/full_fig_p011_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Runtime analysis of the algorithms for the datasets and criteria in Table 2. [PITH_FULL_IMAGE:figures/full_fig_p012_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Relative 𝐿1 error as function of privacy budget for the datasets and criteria from [PITH_FULL_IMAGE:figures/full_fig_p012_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: demonstrates the faithfulness between UMI (Figure 6a) and U 𝐵𝑎𝑦𝑒𝑠 MI (Figure 6b) for the Adult, Stackoverflow survey, and Compas datasets (see Example 1.1) with fairness criteria from [PITH_FULL_IMAGE:figures/full_fig_p016_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Comparison of USAT R (computed by Algorithm 2) with and without the heuristic. (a) Values of U𝑇𝑉 𝐷 MI . (b) Values of UTC [PITH_FULL_IMAGE:figures/full_fig_p027_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Comparison of U𝑇𝑉 𝐷 MI (computed by Algorithm 1) and UTC (computed by Algorithm 3) without privacy consid￾erations. Therefore, [PITH_FULL_IMAGE:figures/full_fig_p027_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Faithfulness of noisy U𝑇𝑉 𝐷 MI (computed by Algorithm 1) to noisy UMI over different datasets and fairness criteria from [PITH_FULL_IMAGE:figures/full_fig_p028_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Effect of 𝑘 on UTC (computed by Algorithm 3) in terms of true value and relative 𝐿1 error for each dataset with its four criteria [PITH_FULL_IMAGE:figures/full_fig_p028_10.png] view at source ↗
read the original abstract

Differential privacy (DP) has become the de facto standard for protecting sensitive data, providing strong guarantees that published statistics or models reveal limited information about any individual. However, privacy noise and restricted data access make it increasingly difficult to assess the fairness and reliability of private datasets. In this paper, we propose a formal framework for quantifying data unfairness under DP. We identify three core desiderata for unfairness measures based on previous work: positivity, monotonicity, and DP computability. We further instantiate them through three complementary measures: (1) a mutual information-based measure with a total variation distance proxy suitable for DP, (2) a data repair-based measure approximated via a reduction to weighted MaxSAT, and (3) a top-$k$ tuple contribution measure that isolates the most influential records in fairness violations. We design privacy-preserving algorithms and analyze their sensitivity, accuracy, and efficiency. Extensive experiments on multiple real-world datasets demonstrate that our proposed measures faithfully approximate their non-private counterparts, effectively quantify bias under privacy constraints, and provide insights for data management.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 2 minor

Summary. The paper proposes a formal framework for quantifying data unfairness under differential privacy. It identifies three core desiderata (positivity, monotonicity, and DP computability) from prior work and instantiates three complementary measures: (1) a mutual information-based measure using a total variation distance proxy, (2) a data repair-based measure reduced to weighted MaxSAT, and (3) a top-k tuple contribution measure. Privacy-preserving algorithms are designed and analyzed for sensitivity, accuracy, and efficiency; experiments on real-world datasets are used to show that the measures faithfully approximate their non-private counterparts and quantify bias under privacy constraints.

Significance. If the measures satisfy the stated properties and the algorithms deliver the claimed accuracy, this could provide useful tools for assessing fairness in differentially private data releases, an area of growing importance at the intersection of databases and privacy. The complementary nature of the three measures and the explicit sensitivity/accuracy analyses are strengths that support practical adoption. The experimental validation of approximation quality is a positive empirical check on the constructions.

major comments (1)
  1. [Framework section] The framework section: the paper adopts the three desiderata (positivity, monotonicity, DP computability) from previous work as sufficient to define unfairness measures under DP, but provides no formal argument, completeness proof, or counterexample analysis showing that these properties capture the relevant notion of database unfairness (as opposed to other possible fairness notions). This is load-bearing for the central claim that the instantiated measures quantify unfairness.
minor comments (2)
  1. [Abstract] The abstract states that experiments demonstrate faithful approximation but does not name the specific real-world datasets used; this information should appear in the experimental section for reproducibility.
  2. [Definitions and algorithms] Notation for the three measures and their DP variants should be unified (e.g., consistent use of subscripts or superscripts) to improve readability across definitions and algorithms.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the constructive feedback and positive assessment of the paper's significance, complementary measures, and experimental validation. We address the major comment below.

read point-by-point responses

  1. Referee: [Framework section] The framework section: the paper adopts the three desiderata (positivity, monotonicity, and DP computability) from previous work as sufficient to define unfairness measures under DP, but provides no formal argument, completeness proof, or counterexample analysis showing that these properties capture the relevant notion of database unfairness (as opposed to other possible fairness notions). This is load-bearing for the central claim that the instantiated measures quantify unfairness.

    Authors: We acknowledge that the framework section adopts the three desiderata from prior work without supplying a formal completeness proof or exhaustive counterexample analysis. These properties are presented as necessary conditions drawn from the literature: positivity to ensure non-negative scores, monotonicity to ensure the measure increases with added bias, and DP computability to permit private estimation. The manuscript does not claim they form a complete axiomatization of all fairness notions. In the revision we will expand the framework section with additional discussion of the rationale for selecting these properties, their relation to dependency-based unfairness, and a brief illustrative example highlighting both their strengths and limitations relative to other possible fairness concepts. This is a partial revision, as a full formal completeness result lies outside the paper's scope. revision: partial

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The paper proposes a framework instantiated via three measures constructed to satisfy desiderata (positivity, monotonicity, DP computability) drawn from prior work, then analyzes algorithm sensitivity/accuracy/efficiency and validates via experiments on real datasets. No equations, self-referential definitions, fitted parameters renamed as predictions, or load-bearing self-citation chains appear in the provided text; the constructions are presented as independent and externally checkable through the stated properties and empirical results.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Abstract-only review; the ledger is limited to the three desiderata explicitly named as foundational. No free parameters, invented entities, or additional axioms are visible.

axioms (1)
  • domain assumption Positivity, monotonicity, and DP computability are the core desiderata for unfairness measures under differential privacy.
    Stated in the abstract as identified from previous work and used to instantiate the three measures.

pith-pipeline@v0.9.0 · 5713 in / 1218 out tokens · 25463 ms · 2026-05-25T05:26:18.793536+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

101 extracted references · 101 canonical work pages · 3 internal anchors

  1. [1]

    Goodfellow, H

    Martín Abadi, Andy Chu, Ian J. Goodfellow, H. Brendan McMahan, Ilya Mironov, Kunal Talwar, and Li Zhang. 2016. Deep Learning with Differential Privacy. In Proceedings of the 2016 ACM SIGSAC Conference on Computer and Communi- cations Security, Vienna, Austria, October 24-28, 2016, Edgar R. Weippl, Stefan Katzenbeisser, Christopher Kruegel, Andrew C. Myers...

    work page doi:10.1145/2976749.2978318 2016

  2. [2]

    John M. Abowd. 2018. The U.S. Census Bureau Adopts Differential Privacy. In Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining(London, United Kingdom)(KDD ’18). Association for Computing Machinery, New York, NY, USA, 2867–2867. https://doi.org/10.1145/ 3219819.3226070

    work page arXiv 2018

  3. [3]

    Afrati and Phokion G

    Foto N. Afrati and Phokion G. Kolaitis. 2009. Repair Checking in Inconsistent Databases: Algorithms and Complexity. InICDT. 31–41

    work page 2009

  4. [4]

    Alekh Agarwal, Alina Beygelzimer, Miroslav Dudík, John Langford, and Hanna M. Wallach. 2018. A Reductions Approach to Fair Classification. InProceedings of the 35th International Conference on Machine Learning, ICML 2018, Stock- holmsmässan, Stockholm, Sweden, July 10-15, 2018 (Proceedings of Machine Learn- ing Research), Jennifer G. Dy and Andreas Krause...

    work page 2018

  5. [5]

    Alon Albalak, Yanai Elazar, Sang Michael Xie, Shayne Longpre, Nathan Lambert, Xinyi Wang, Niklas Muennighoff, Bairu Hou, Liangming Pan, Haewon Jeong, Colin Raffel, Shiyu Chang, Tatsunori Hashimoto, and William Yang Wang. 2024. A Survey on Data Selection for Language Models.Trans. Mach. Learn. Res.2024 (2024). https://openreview.net/forum?id=XfHWcNTSHp

    work page 2024

  6. [6]

    Sergül Aydöre, William Brown, Michael Kearns, Krishnaram Kenthapadi, Luca Melis, Aaron Roth, and Amaresh Ankit Siva. 2021. Differentially Private Query Release Through Adaptive Projection. InICML, Vol. 139. 457–467

    work page 2021

  7. [7]

    2023.Fairness and machine learning: Limitations and opportunities

    Solon Barocas, Moritz Hardt, and Arvind Narayanan. 2023.Fairness and machine learning: Limitations and opportunities. MIT Press

    work page 2023

  8. [8]

    Richard Berk, Hoda Heidari, Shahin Jabbari, Michael Kearns, and Aaron Roth

    work page

  9. [9]

    Fairness in Criminal Justice Risk Assessments: The State of the Art.Socio- logical Methods & Research50, 1 (2021), 3–44

    work page 2021

  10. [10]

    Bertossi

    Leopoldo E. Bertossi. 2018. Measuring and Computing Database Inconsistency via Repairs. InScalable Uncertainty Management - 12th International Conference, SUM 2018, Milan, Italy, October 3-5, 2018, Proceedings (Lecture Notes in Computer Science), Davide Ciucci, Gabriella Pasi, and Barbara Vantaggi (Eds.), Vol. 11142. Springer, 368–372. https://doi.org/10....

    work page doi:10.1007/978-3-030-00461-3_26 2018

  11. [11]

    Bertossi, Solmaz Kolahi, and Laks V

    Leopoldo E. Bertossi, Solmaz Kolahi, and Laks V. S. Lakshmanan. 2013. Data Cleaning and Query Answering with Matching Dependencies and Matching Functions.Theory Comput. Syst.52, 3 (2013), 441–482

    work page 2013

  12. [12]

    Rob Brennan, Judie Attard, Plamen Petkov, Tadhg Nagle, and Markus Helfert

    work page

  13. [13]

    Exploring Data Value Assessment: A Survey Method and Investigation of the Perceived Relative Importance of Data Value Dimensions. InProceedings of the 21st International Conference on Enterprise Information Systems, ICEIS 2019, Heraklion, Crete, Greece, May 3-5, 2019, Volume 1, Joaquim Filipe, Michal Smialek, Alexander Brodsky, and Slimane Hammoudi (Eds.)...

    work page doi:10.5220/0007723402000207 2019

  14. [14]

    Marc-Etienne Brunet, Colleen Alkalay-Houlihan, Ashton Anderson, and Richard S. Zemel. 2019. Understanding the Origins of Bias in Word Embed- dings. InProceedings of the 36th International Conference on Machine Learning, ICML 2019, 9-15 June 2019, Long Beach, California, USA (Proceedings of Machine Learning Research), Kamalika Chaudhuri and Ruslan Salakhut...

    work page 2019

  15. [15]

    Toon Calders, Faisal Kamiran, and Mykola Pechenizkiy. 2009. Building Classifiers with Independency Constraints. InICDM Workshops 2009, IEEE International Conference on Data Mining Workshops, Miami, Florida, USA, 6 December 2009, Yücel Saygin, Jeffrey Xu Yu, Hillol Kargupta, Wei Wang, Sanjay Ranka, Philip S. Yu, and Xindong Wu (Eds.). IEEE Computer Society...

    work page 2009

  16. [16]

    Calmon, Dennis Wei, Bhanukiran Vinzamuri, Karthikeyan Natesan Ramamurthy, and Kush R

    Flávio P. Calmon, Dennis Wei, Bhanukiran Vinzamuri, Karthikeyan Natesan Ramamurthy, and Kush R. Varshney. 2017. Optimized Pre-Processing for Dis- crimination Prevention. InAdvances in Neural Information Processing Systems 30: Annual Conference on Neural Information Processing Systems 2017, Decem- ber 4-9, 2017, Long Beach, CA, USA, Isabelle Guyon, Ulrike ...

    work page 2017

  17. [17]

    California Consumer Privacy Act (CCPA)

    CCPA 2023. California Consumer Privacy Act (CCPA). https://oag.ca.gov/privacy/ccpa

    work page 2023

  18. [18]

    Dingfan Chen, Tribhuvanesh Orekondy, and Mario Fritz. 2020. GS-WGAN: A Gradient-Sanitized Approach for Learning Differentially Private Generators. In NIPS

    work page 2020

  19. [19]

    Alexandra Chouldechova. 2017. Fair Prediction with Disparate Impact: A Study of Bias in Recidivism Prediction Instruments.Big Data5, 2 (2017), 153–163

    work page 2017

  20. [20]

    Ilyas, and Paolo Papotti

    Xu Chu, Ihab F. Ilyas, and Paolo Papotti. 2013. Holistic data cleaning: Putting violations into context. InICDE. 458–469

    work page 2013

  21. [21]

    Sam Corbett-Davies, Emma Pierson, Avi Feller, Sharad Goel, and Aziz Huq

    work page

  22. [22]

    InProceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, Halifax, NS, Canada, August 13 - 17, 2017

    Algorithmic Decision Making and the Cost of Fairness. InProceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, Halifax, NS, Canada, August 13 - 17, 2017. ACM, 797–806. https: //doi.org/10.1145/3097983.3098095

    work page doi:10.1145/3097983.3098095 2017

  23. [23]

    T. M. Cover and Joy A. Thomas. 2005.Elements of information theory. Wiley- Interscience

    work page 2005

  24. [24]

    2011.Information theory: coding theorems for discrete memoryless systems

    Imre Csiszár and János Körner. 2011.Information theory: coding theorems for discrete memoryless systems. Cambridge University Press

    work page 2011

  25. [25]

    Daniel Deutch, Nave Frost, Amir Gilad, and Oren Sheffer. 2021. Explanations for Data Repair Through Shapley Values. InCIKM. 362–371

    work page 2021

  26. [26]

    Bolin Ding, Janardhan Kulkarni, and Sergey Yekhanin. 2017. Collecting Telemetry Data Privately. InProceedings of the 31st International Conference on Neural Information Processing Systems(Long Beach, California, USA)(NIPS’17). Curran Associates Inc., Red Hook, NY, USA, 3574–3583

    work page 2017

  27. [27]

    Wei Dong, Juanru Fang, Ke Yi, Yuchao Tao, and Ashwin Machanavajjhala. 2022. R2t: Instance-optimal truncation for differentially private query evaluation with foreign keys. InSIGMOD. 759–772

    work page 2022

  28. [28]

    Cynthia Dwork. 2006. Differential Privacy. InAutomata, Languages and Pro- gramming, Michele Bugliesi, Bart Preneel, Vladimiro Sassone, and Ingo Wegener (Eds.). Springer Berlin Heidelberg

    work page 2006

  29. [29]

    Cynthia Dwork, Moritz Hardt, Toniann Pitassi, Omer Reingold, and Richard S. Zemel. 2012. Fairness through awareness. InInnovations in Theoretical Computer Science 2012, Cambridge, MA, USA, January 8-10, 2012. 214–226

    work page 2012

  30. [30]

    Cynthia Dwork, Frank McSherry, Kobbi Nissim, and Adam Smith. 2006. Cali- brating Noise to Sensitivity in Private Data Analysis. InProceedings of the Third Conference on Theory of Cryptography(New York, NY)(TCC’06). Springer-Verlag, Berlin, Heidelberg, 265–284. https://doi.org/10.1007/11681878_14

    work page doi:10.1007/11681878_14 2006

  31. [31]

    Cynthia Dwork and Aaron Roth. 2014. The Algorithmic Foundations of Differ- ential Privacy.Found. Trends Theor. Comput. Sci.(2014)

    work page 2014

  32. [32]

    Úlfar Erlingsson, Vasyl Pihur, and Aleksandra Korolova. 2014. RAPPOR: Ran- domized Aggregatable Privacy-Preserving Ordinal Response. InProceedings of the 2014 ACM SIGSAC Conference on Computer and Communications Security (Scottsdale, Arizona, USA)(CCS ’14). Association for Computing Machinery, New York, NY, USA, 1054–1067. https://doi.org/10.1145/2660267.2660348

    work page doi:10.1145/2660267.2660348 2014

  33. [33]

    Anna Fariha and Alexandra Meliou. 2019. Example-driven query intent discovery: Abductive reasoning using semantic similarity.arXiv preprint arXiv:1906.10322 (2019)

    work page internal anchor Pith review Pith/arXiv arXiv 2019

  34. [34]

    Friedler, John Moeller, Carlos Scheidegger, and Suresh Venkatasubramanian

    Michael Feldman, Sorelle A. Friedler, John Moeller, Carlos Scheidegger, and Suresh Venkatasubramanian. 2015. Certifying and Removing Disparate Impact. InProceedings of the 21th ACM SIGKDD International Conference on Knowl- edge Discovery and Data Mining, Sydney, NSW, Australia, August 10-13, 2015, Longbing Cao, Chengqi Zhang, Thorsten Joachims, Geoffrey I...

    work page doi:10.1145/2783258.2783311 2015

  35. [35]

    Robert Warren, and Michael Westberry

    Sarah Flood, Miriam King, Renae Rodgers, Steven Ruggles, J. Robert Warren, and Michael Westberry. 2021. Integrated Public Use Microdata Series, Current Population Survey: Version 9.0 [dataset].Minneapolis, MN: IPUMS(2021). https: //doi.org/10.18128/D030.V9.0

    work page doi:10.18128/d030.v9.0 2021

  36. [36]

    Maurice Fréchet. 1935. Généralisation du théoreme des probabilités totales. Fundamenta mathematicae25, 1 (1935), 379–387

    work page 1935

  37. [37]

    Sainyam Galhotra, Yuriy Brun, and Alexandra Meliou. 2017. Fairness testing: testing software for discrimination. InESEC/FSE. 498–510

    work page 2017

  38. [38]

    2016-04-27. Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016 on the protection of natural persons with regard to the processing of personal data and on the free movement of such data, and repealing Directive 95/46/EC (General Data Protection Regulation).OJ(2016-04-27)

    work page 2016

  39. [39]

    Chang Ge, Shubhankar Mohapatra, Xi He, and Ihab F. Ilyas. 2021. Kamino: Constraint-Aware Differentially Private Data Synthesis.Proc. VLDB Endow.14, 10 (2021), 1886–1899

    work page 2021

  40. [40]

    Floris Geerts, Giansalvatore Mecca, Paolo Papotti, and Donatello Santoro. 2013. The LLUNATIC Data-Cleaning Framework.Proc. VLDB Endow.6, 9 (2013), 625– 636

    work page 2013

  41. [41]

    Arpita Ghosh and Aaron Roth. 2011. Selling privacy at auction. InProceedings 12th ACM Conference on Electronic Commerce (EC-2011), San Jose, CA, USA, June 5-9, 2011, Yoav Shoham, Yan Chen, and Tim Roughgarden (Eds.). ACM, 199–208. https://doi.org/10.1145/1993574.1993605

    work page doi:10.1145/1993574.1993605 2011

  42. [42]

    Amir Gilad, Daniel Deutch, and Sudeepa Roy. 2020. On Multiple Semantics for Declarative Database Repairs. InSIGMOD. 817–831

    work page 2020

  43. [43]

    John Grant and Anthony Hunter. 2006. Measuring inconsistency in knowledge- bases.J. Intell. Inf. Syst.27, 2 (2006), 159–184. https://doi.org/10.1007/S10844- 006-2974-4

    work page doi:10.1007/s10844- 2006

  44. [44]

    John Grant and Anthony Hunter. 2017. Analysing inconsistent information using distance-based measures.Int. J. Approx. Reason.89 (2017), 3–26. https: //doi.org/10.1016/J.IJAR.2016.04.004

    work page doi:10.1016/j.ijar.2016.04.004 2017

  45. [45]

    Moritz Hardt, Katrina Ligett, and Frank Mcsherry. [n.d.]. A Simple and Practical Algorithm for Differentially Private Data Release. InNIPS. Curran Associates, Inc., 2339–2347

    work page

  46. [46]

    Moritz Hardt, Eric Price, Eric Price, and Nati Srebro. 2016. Equality of Opportunity in Supervised Learning. InNIPS, D. Lee, M. Sugiyama, U. Luxburg, I. Guyon, and R. Garnett (Eds.), Vol. 29

    work page 2016

  47. [47]

    Rothblum

    Moritz Hardt and Guy N. Rothblum. 2010. A Multiplicative Weights Mechanism for Privacy-Preserving Data Analysis. InFOCS. 61–70

    work page 2010

  48. [48]

    Alireza Heidari, Joshua McGrath, Ihab F Ilyas, and Theodoros Rekatsinas. 2019. Holodetect: Few-shot learning for error detection. InProceedings of the 2019 International Conference on Management of Data. 829–846

    work page 2019

  49. [49]

    Hoda Heidari, Michele Loi, Krishna P Gummadi, and Andreas Krause. 2019. A moral framework for understanding fair ml through economic models of equality of opportunity. InProceedings of the conference on fairness, accountability, and transparency. 181–190

    work page 2019

  50. [50]

    Anthony Hunter and Sébastien Konieczny. 2010. On the measure of conflicts: Shapley Inconsistency Values.Artif. Intell.174, 14 (2010), 1007–1026. https: //doi.org/10.1016/J.ARTINT.2010.06.001

    work page doi:10.1016/j.artint.2010.06.001 2010

  51. [51]

    Bargav Jayaraman and David Evans. 2019. Evaluating Differentially Private Machine Learning in Practice. In28th USENIX Security Symposium, USENIX Security 2019, Santa Clara, CA, USA, August 14-16, 2019, Nadia Heninger and Patrick Traynor (Eds.). USENIX Association, 1895–1912. https://www.usenix. org/conference/usenixsecurity19/presentation/jayaraman

    work page 2019

  52. [52]

    Faisal Kamiran and Toon Calders. 2011. Data preprocessing techniques for classification without discrimination.Knowl. Inf. Syst.33, 1 (2011), 1–33. https: //doi.org/10.1007/S10115-011-0463-8

    work page doi:10.1007/s10115-011-0463-8 2011

  53. [53]

    Maurice G Kendall. 1938. A new measure of rank correlation.Biometrika30, 1-2 (1938), 81–93

    work page 1938

  54. [54]

    Nohyun Ki, Hoyong Choi, and Hye Won Chung. 2023. Data Valuation With- out Training of a Model. InThe Eleventh International Conference on Learn- ing Representations, ICLR 2023, Kigali, Rwanda, May 1-5, 2023. OpenReview.net. https://openreview.net/forum?id=XIzO8zr-WbM

    work page 2023

  55. [55]

    Niki Kilbertus, Mateo Rojas-Carulla, Giambattista Parascandolo, Moritz Hardt, Dominik Janzing, and Bernhard Schölkopf. 2017. Avoiding Discrimination through Causal Reasoning. InNIPS. 656–666

    work page 2017

  56. [56]

    Kleinberg, Sendhil Mullainathan, and Manish Raghavan

    Jon M. Kleinberg, Sendhil Mullainathan, and Manish Raghavan. 2017. Inherent Trade-Offs in the Fair Determination of Risk Scores. InITCS, Vol. 67. 43:1–43:23

    work page 2017

  57. [57]

    Kevin M. Knight. 2003. Two Information Measures for Inconsistent Sets.J. Log. Lang. Inf.12, 2 (2003), 227–248. https://doi.org/10.1023/A:1022351919320

    work page doi:10.1023/a:1022351919320 2003

  58. [58]

    Sébastien Konieczny, Jérôme Lang, and Pierre Marquis. 2003. Quantifying infor- mation and contradiction in propositional logic through test actions. InIJCAI-03, Proceedings of the Eighteenth International Joint Conference on Artificial Intelli- gence, Acapulco, Mexico, August 9-15, 2003, Georg Gottlob and Toby Walsh (Eds.). Morgan Kaufmann, 106–111. http:...

    work page 2003

  59. [59]

    Ios Kotsogiannis, Yuchao Tao, Xi He, Maryam Fanaeepour, Ashwin Machanava- jjhala, Michael Hay, and Gerome Miklau. 2019. PrivateSQL: A Differentially Private SQL Query Engine.Proc. VLDB Endow.12, 11 (2019), 1371–1384

    work page 2019

  60. [60]

    Alexander Kraskov, Harald Stögbauer, and Peter Grassberger. 2004. Estimating mutual information.Physical Review E—Statistical, Nonlinear, and Soft Matter Physics69, 6 (2004), 066138

    work page 2004

  61. [61]

    Kusner, Joshua R

    Matt J. Kusner, Joshua R. Loftus, Chris Russell, and Ricardo Silva. 2017. Counter- factual Fairness. InNIPS. 4066–4076

    work page 2017

  62. [62]

    Chao Li, Daniel Yang Li, Gerome Miklau, and Dan Suciu. 2017. A theory of pricing private data.Commun. ACM60, 12 (2017), 79–86. https://doi.org/10.1145/3139457

    work page doi:10.1145/3139457 2017

  63. [63]

    Chao Li and Gerome Miklau. 2013. Optimal error of query sets under the differentially-private matrix mechanism. InICDT. 272–283

    work page 2013

  64. [64]

    Ninghui Li, Zhikun Zhang, and Tianhao Wang. 2021. DPSyn: Experiences in the NIST Differential Privacy Data Synthesis Challenges.CoRRabs/2106.12949 (2021)

    work page arXiv 2021

  65. [65]

    Ullman, and Zhi- wei Steven Wu

    Terrance Liu, Giuseppe Vietri, Thomas Steinke, Jonathan R. Ullman, and Zhi- wei Steven Wu. 2021. Leveraging Public Data for Practical Private Query Release. InICML, Vol. 139. 6968–6977

    work page 2021

  66. [66]

    Bertossi, Benny Kimelfeld, and Moshe Sebag

    Ester Livshits, Leopoldo E. Bertossi, Benny Kimelfeld, and Moshe Sebag. 2021. The Shapley Value of Tuples in Query Answering.Log. Methods Comput. Sci.17, 3 (2021). https://doi.org/10.46298/LMCS-17(3:22)2021

    work page doi:10.46298/lmcs-17(3:22)2021 2021

  67. [67]

    Ester Livshits and Benny Kimelfeld. 2021. The Shapley Value of Inconsistency Measures for Functional Dependencies. InICDT, Vol. 186. 15:1–15:19

    work page 2021

  68. [68]

    Ester Livshits, Benny Kimelfeld, and Sudeepa Roy. 2020. Computing Optimal Repairs for Functional Dependencies.ACM Trans. Database Syst.45, 1 (2020), 4:1–4:46. https://doi.org/10.1145/3360904

    work page doi:10.1145/3360904 2020

  69. [69]

    Ilyas, Benny Kimelfeld, and Sudeepa Roy

    Ester Livshits, Rina Kochirgan, Segev Tsur, Ihab F. Ilyas, Benny Kimelfeld, and Sudeepa Roy. 2021. Properties of Inconsistency Measures for Databases. In SIGMOD. 1182–1194

    work page 2021

  70. [70]

    Simari, V

    Maria Vanina Martinez, Andrea Pugliese, Gerardo I. Simari, V. S. Subrahmanian, and Henri Prade. 2007. How Dirty Is Your Relational Database? An Axiomatic Approach. InSymbolic and Quantitative Approaches to Reasoning with Uncertainty, 9th European Conference, ECSQARU 2007, Hammamet, Tunisia, October 31 - No- vember 2, 2007, Proceedings (Lecture Notes in Co...

    work page doi:10.1007/978-3-540-75256-1_12 2007

  71. [71]

    Ryan McKenna, Gerome Miklau, Michael Hay, and Ashwin Machanavajjhala

    work page

  72. [72]

    Optimizing Error of High-dimensional Statistical Queries Under Differen- tial Privacy.PVLDB11, 10 (2018)

    work page 2018

  73. [73]

    Ryan McKenna, Gerome Miklau, and Daniel Sheldon. 2021. Winning the NIST Contest: A scalable and general approach to differentially private synthetic data. CoRRabs/2108.04978 (2021)

    work page arXiv 2021

  74. [74]

    Moore, and Dan Suciu

    Alexandra Meliou, Wolfgang Gatterbauer, Katherine F. Moore, and Dan Suciu

    work page

  75. [75]

    VLDB Endow.4, 1 (2010), 34–45

    The Complexity of Causality and Responsibility for Query Answers and non-Answers.Proc. VLDB Endow.4, 1 (2010), 34–45

    work page 2010

  76. [76]

    Shubhankar Mohapatra, Amir Gilad, Xi He, and Benny Kimelfeld. 2025. Com- puting Inconsistency Measures Under Differential Privacy.Proc. ACM Manag. Data3, 3 (2025), 140:1–140:27. https://doi.org/10.1145/3725397

    work page doi:10.1145/3725397 2025

  77. [77]

    Kedian Mu, Weiru Liu, and Zhi Jin. 2011. A general framework for measuring inconsistency through minimal inconsistent sets.Knowl. Inf. Syst.27, 1 (2011), 85–114. https://doi.org/10.1007/S10115-010-0295-Y

    work page doi:10.1007/s10115-010-0295-y 2011

  78. [78]

    Razieh Nabi and Ilya Shpitser. 2018. Fair Inference on Outcomes. InAAAI. 1931–1940

    work page 2018

  79. [79]

    Near and Xi He

    Joseph P. Near and Xi He. 2021. Differential Privacy for Databases.Found. Trends Databases11, 2 (2021), 109–225. https://doi.org/10.1561/1900000066

    work page doi:10.1561/1900000066 2021

  80. [80]

    Goodfellow, and Kunal Talwar

    Nicolas Papernot, Martín Abadi, Úlfar Erlingsson, Ian J. Goodfellow, and Kunal Talwar. 2017. Semi-supervised Knowledge Transfer for Deep Learning from Private Training Data. In5th International Conference on Learning Representa- tions, ICLR 2017, Toulon, France, April 24-26, 2017, Conference Track Proceedings. OpenReview.net. https://openreview.net/forum?...

    work page 2017

Showing first 80 references.