mlsanityData to 7 Oct 2026

Fairness-Aware PAC Learning from Corrupted Data

arXiv 2102.06004 ↗First posted 11 Feb 2021cs.LG stat.MLcited by 18

Not on the map. Why?

Abstract (arXiv)

Addressing fairness concerns about machine learning models is a crucial step towards their long-term adoption in real-world automated systems. While many approaches have been developed for training fair models from data, little is known about the robustness of these methods to data corruption. In this work we consider fairness-aware learning under worst-case data manipulations. We show that an adversary can in some situations force any learner to return an overly biased classifier, regardless of the sample size and with or without degrading accuracy, and that the strength of the excess bias increases for learning problems with underrepresented protected groups in the data. We also prove that our hardness results are tight up to constant factors. To this end, we study two natural learning algorithms that optimize for both accuracy and fairness and show that these algorithms enjoy guarantees that are order-optimal in terms of the corruption ratio and the protected groups frequencies in the large data limit.

Key facts. Cited by 18 papers in the map (data through 7 Oct 2026). Builds on 34 of its 54 references that are in the map. First posted on arXiv 11 Feb 2021.

Builds on

34 of 54 references.

References · 34

  1. 2016Equality of Opportunity in Supervised Learningcited by 2,000
  2. 2019A Survey on Bias and Fairness in Machine Learningcited by 1,911
  3. 2017Targeted Backdoor Attacks on Deep Learning Systems Using Data Poisoningcited by 1,012
  4. 2018A Reductions Approach to Fair Classificationcited by 585
  5. 2016Inherent Trade-Offs in the Fair Determination of Risk Scorescited by 582
  6. 2017Wild Patterns: Ten Years After the Rise of Adversarial Machine Learningcited by 411
  7. 2018Fairness Without Demographics in Repeated Loss Minimizationcited by 363
  8. 2017Certified Defenses for Data Poisoning Attackscited by 288
  9. 2020Fairness without Demographics through Adversarially Reweighted Learningcited by 251
  10. 2018Using Trusted Data to Train Deep Networks on Labels Corrupted by Severe Noisecited by 216
  11. 2016Robust Estimators in High Dimensions without the Computational Intractabilitycited by 209
  12. 2019Identifying and Correcting Label Bias in Machine Learningcited by 146
22 more
  1. 2018Sever: A Robust Meta-Algorithm for Stochastic Optimizationcited by 123
  2. 2020Tilted Empirical Risk Minimizationcited by 103
  3. 2016Learning from Untrusted Datacited by 90
  4. 2019Fairness risk measurescited by 85
  5. 2019Assessing Algorithmic Fairness with Unobserved Protected Class Using Data Combinationcited by 79
  6. 2020Robust Optimization for Fairness with Noisy Protected Groupscited by 78
  7. 2020Fair Classification with Group-Dependent Label Noisecited by 67
  8. 2020Poisoning Attacks on Algorithmic Fairnesscited by 58
  9. 2020FR-Train: A Mutual Information-Based Approach to Fair and Robust Trainingcited by 52
  10. 2019Recovering from Biased Data: Can Fairness Constraints Improve Accuracy?cited by 50
  11. 2018Residual Unfairness in Fair Machine Learning from Prejudiced Datacited by 49
  12. 2019Equalized odds postprocessing under imperfect group informationcited by 48
  13. 2020Exacerbating Algorithmic Bias through Fairness Attackscited by 47
  14. 2020Fair Learning with Private Demographic Datacited by 46
  15. 2020A Distributionally Robust Approach to Fair Classificationcited by 43
  16. 2019Noise-tolerant fair classificationcited by 40
  17. 2020Fairness Evaluation in Presence of Biased Noisy Labelscited by 38
  18. 2020On Adversarial Bias and the Robustness of Fair Machine Learningcited by 36
  19. 2020Mitigating Bias in Set Selection with Noisy Protected Attributescited by 29
  20. 2019Robust Learning from Untrusted Sourcescited by 27
  21. 2018Learning under selective labels in the presence of expert consistencycited by 18
  22. 2019Data Cleaning for Accurate, Fair, and Robust Models: A Big Data - AI Integration Approachcited by 10

Built on by

Cited by year

Citing papers per year: 2022: 4; 2023: 6; 2024: 5; 2025: 3.

18 papers.

Sort:

By area:

  1. 2022Towards Adversarial Evaluations for Inexact Machine Unlearningcited by 82
  2. 2024Corrective Machine Unlearningcited by 26
  3. 2023Monitoring Algorithmic Fairnesscited by 8
  4. 2024The Fragility of Fairness: Causal Sensitivity Analysis for Fair Machine Learningcited by 8
  5. 2023Monitoring Algorithmic Fairness under Partial Observationscited by 7
  6. 2023Quantifying and mitigating the impact of label errors on model disparity metricscited by 7
  7. 2022Breaking Fair Binary Classification with Optimal Flipping Attackscited by 4
  8. 2022Generalization In Multi-Objective Machine Learningcited by 4
  9. 2023Mitigating Source Bias for Fairer Weak Supervisioncited by 3
  10. 2022A Survey on Preserving Fairness Guarantees in Changing Environmentscited by 2
  11. 2024A Cognac Shot To Forget Bad Memories: Corrective Unlearning for Graph Neural Networkscited by 2
  12. 2024Navigating Towards Fairness with Data Selectioncited by 1
6 more
  1. 2023Measuring Equality in Machine Learning Security Defenses: A Case Study in Speech Recognitioncited by 0
  2. 2023Identifying Early Help Referrals For Local Authorities With Machine Learning And Bias Analysiscited by 0
  3. 2024Utility of Multimodal Large Language Models in Analyzing Chest X-ray with Incomplete Contextual Informationcited by 0
  4. 2025A Critical Review of Predominant Bias in Neural Networkscited by 0
  5. 2025Monitoring of Static Fairnesscited by 0
  6. 2025Algorithmic Fairness: A Runtime Perspectivecited by 0