mlsanityData to 7 Oct 2026

A Unified View of Group Fairness Tradeoffs Using Partial Information Decomposition

arXiv 2406.04562 ↗First posted 7 Jun 2024cs.IT cs.CY cs.LGcited by 3

Not on the map. Why?

Abstract (arXiv)

This paper introduces a novel information-theoretic perspective on the relationship between prominent group fairness notions in machine learning, namely statistical parity, equalized odds, and predictive parity. It is well known that simultaneous satisfiability of these three fairness notions is usually impossible, motivating practitioners to resort to approximate fairness solutions rather than stringent satisfiability of these definitions. However, a comprehensive analysis of their interrelations, particularly when they are not exactly satisfied, remains largely unexplored. Our main contribution lies in elucidating an exact relationship between these three measures of (un)fairness by leveraging a body of work in information theory called partial information decomposition (PID). In this work, we leverage PID to identify the granular regions where these three measures of (un)fairness overlap and where they disagree with each other leading to potential tradeoffs. We also include numerical simulations to complement our results.

Key facts. Cited by 3 papers in the map (data through 7 Oct 2026). Builds on 31 of its 68 references that are in the map. First posted on arXiv 7 Jun 2024.

Builds on

31 of 68 references.

References · 31

  1. 2016Equality of Opportunity in Supervised Learningcited by 2,000
  2. 2019A Survey on Bias and Fairness in Machine Learningcited by 1,911
  3. 2016Fair prediction with disparate impact: A study of bias in recidivism prediction instrumentscited by 671
  4. 2016Inherent Trade-Offs in the Fair Determination of Risk Scorescited by 582
  5. 2017On Fairness and Calibrationcited by 406
  6. 2019Inherent Tradeoffs in Learning Fair Representationscited by 144
  7. 2018Why Is My Classifier Discriminatory?cited by 143
  8. 2021A Clarification of the Nuances in the Fairness Metrics Landscapecited by 98
  9. 2023Robust Counterfactual Explanations for Neural Networks With Probabilistic Guaranteescited by 45
  10. 2020Accuracy and Fairness Trade-offs in Machine Learning: A Stochastic Multi-Objective Approachcited by 39
  11. 2021Enforcing fairness in private federated learning via the modified method of differential multiplierscited by 38
  12. 2019R'enyi Fair Inferencecited by 33
19 more
  1. 2022Active Fairness Auditingcited by 27
  2. 2020Causal Feature Selection for Algorithmic Fairnesscited by 25
  3. 2021A Rigorous Information-Theoretic Definition of Redundancy and Relevancy in Feature Selection Based on (Partial) Information Decompositioncited by 25
  4. 2023Aleatoric and Epistemic Discrimination: Fundamental Limits of Fairness Interventionscited by 19
  5. 2023The Possibility of Fairness: Revisiting the Impossibility Theorem in Practicecited by 18
  6. 2018Fairness in Supervised Learning: An Information Theoretic Approachcited by 17
  7. 2023Causal Context Connects Counterfactual Fairness to Robust Prediction and Group Fairnesscited by 17
  8. 2019Fairness-Aware Neural R'eyni Minimization for Continuous Featurescited by 15
  9. 2020Fairness Under Feature Exemptions: Counterfactual and Observational Measurescited by 15
  10. 2022Pushing the limits of fairness impossibility: Who's the fairest of them all?cited by 15
  11. 2023Gaussian Partial Information Decomposition: Bias Correction and Application to High-dimensional Datacited by 15
  12. 2019Generating Fair Universal Representations using Adversarial Modelscited by 12
  13. 2023More Synergy, Less Redundancy: Exploiting Joint Mutual Information for Self-Supervised Learningcited by 11
  14. 2023Demystifying Local and Global Fairness Trade-offs in Federated Learning Using Partial Information Decompositioncited by 10
  15. 2017The cost of fairness in classificationcited by 9
  16. 2021Gradual (In)Compatibility of Fairness Criteriacited by 8
  17. 2022Can Querying for Bias Leak Protected Attributes? Achieving Privacy With Smooth Sensitivitycited by 8
  18. 2024Intrinsic Fairness-Accuracy Tradeoffs under Equalized Oddscited by 7
  19. 2021Can Information Flows Suggest Targets for Interventions in Neural Circuits?cited by 6

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 1; 2025: 1; 2026: 1.

3 papers.

  1. 2024Quantifying Knowledge Distillation Using Partial Information Decompositioncited by 12
  2. 2025Fair Classification by Direct Intervention on Operating Characteristicscited by 0
  3. 2026Interpreting Reasoning of Large Language Models via Partial Information Decompositioncited by 0