mlsanityData to 7 Oct 2026

Quantifying and mitigating the impact of label errors on model disparity metrics

arXiv 2310.02533 ↗First posted 4 Oct 2023cs.LG stat.MLcited by 7

Not on the map. Why?

Abstract (arXiv)

Errors in labels obtained via human annotation adversely affect a model's performance. Existing approaches propose ways to mitigate the effect of label error on a model's downstream accuracy, yet little is known about its impact on a model's disparity metrics. Here we study the effect of label error on a model's disparity metrics. We empirically characterize how varying levels of label error, in both training and test data, affect these disparity metrics. We find that group calibration and other metrics are sensitive to train-time and test-time label error -- particularly for minority groups. This disparate effect persists even for models trained with noise-aware algorithms. To mitigate the impact of training-time label error, we present an approach to estimate the influence of a training input's label on a model's group disparity metric. We empirically assess the proposed approach on a variety of datasets and find significant improvement, compared to alternative approaches, in identifying training inputs that improve a model's disparity metric. We complement the approach with an automatic relabel-and-finetune scheme that produces updated models with, provably, improved group calibration error.

Key facts. Cited by 7 papers in the map (data through 7 Oct 2026). Builds on 48 of its 73 references that are in the map. First posted on arXiv 4 Oct 2023.

Builds on

48 of 73 references.

References · 48

  1. 2017mixup: Beyond Empirical Risk Minimizationcited by 5,155
  2. 2017On Calibration of Modern Neural Networkscited by 4,860
  3. 2019Unsupervised Cross-lingual Representation Learning at Scalecited by 3,907
  4. 2016Equality of Opportunity in Supervised Learningcited by 2,000
  5. 2017Understanding Black-box Predictions via Influence Functionscited by 1,714
  6. 2017A Closer Look at Memorization in Deep Networkscited by 936
  7. 2016Semantics derived automatically from language corpora contain human-like biasescited by 894
  8. 2020DivideMix: Learning with Noisy Labels as Semi-supervised Learningcited by 691
  9. 2018Learning to Reweight Examples for Robust Deep Learningcited by 670
  10. 2020Estimating Training Data Influence by Tracing Gradient Descentcited by 565
  11. 2020What Neural Networks Memorize and Why: Discovering the Long Tail via Influence Estimationcited by 492
  12. 2021Deep Neural Networks and Tabular Data: A Surveycited by 489
36 more
  1. 2020GoEmotions: A Dataset of Fine-Grained Emotionscited by 479
  2. 2021Retiring Adult: New Datasets for Fair Machine Learningcited by 443
  3. 2017Robust Loss Functions under Label Noise for Deep Neural Networkscited by 441
  4. 2019Confident Learning: Estimating Uncertainty in Dataset Labelscited by 411
  5. 2017On Fairness and Calibrationcited by 406
  6. 2021Pervasive Label Errors in Test Sets Destabilize Machine Learning Benchmarkscited by 388
  7. 2014Training Deep Neural Networks on Noisy Labels with Bootstrappingcited by 322
  8. 2019Differential Privacy Has Disparate Impact on Model Accuracycited by 316
  9. 2020Environment Inference for Invariant Learningcited by 306
  10. 2020Combating noisy labels by agreement: A joint training method with co-regularizationcited by 289
  11. 2020Normalized Loss Functions for Deep Learning with Noisy Labelscited by 257
  12. 2018Multiaccuracy: Black-Box Post-Processing for Fairness in Classificationcited by 221
  13. 2018Using Trusted Data to Train Deep Networks on Labels Corrupted by Severe Noisecited by 216
  14. 2018Representer Point Selection for Explaining Deep Neural Networkscited by 187
  15. 2017Deep Learning is Robust to Massive Label Noisecited by 164
  16. 2019What Do Compressed Deep Neural Networks Forget?cited by 164
  17. 2019Identifying and Correcting Label Bias in Machine Learningcited by 146
  18. 2020Explaining Black Box Predictions and Unveiling Data Artifacts through Influence Functionscited by 144
  19. 2020To be Robust or to be Fair: Towards Fairness in Adversarial Trainingcited by 137
  20. 2020Certified Robustness to Label-Flipping Attacks via Randomized Smoothingcited by 103
  21. 2020Multivariate Confidence Calibration for Object Detectioncited by 98
  22. 2018Understanding the Origins of Bias in Word Embeddingscited by 90
  23. 2020Fair Classification with Group-Dependent Label Noisecited by 67
  24. 2022A Systematic Study of Bias Amplificationcited by 65
  25. 2022Instance-Dependent Label-Noise Learning with Manifold-Regularized Transition Matrix Estimationcited by 49
  26. 2022Pruning has a disparate impact on model accuracycited by 43
  27. 2020Fairness Evaluation in Presence of Biased Noisy Labelscited by 38
  28. 2022Achieving Fairness at No Utility Cost via Data Reweighing with Influencecited by 38
  29. 2020Selective Classification Can Magnify Disparities Across Groupscited by 35
  30. 2021Interpretable Data-Based Explanations for Fairness Debuggingcited by 30
  31. 2022Fair Infinitesimal Jackknife: Mitigating the Influence of Biased Training Data Points Without Refittingcited by 26
  32. 2021Fairness On The Ground: Applying Algorithmic Fairness Approaches to Production Systemscited by 20
  33. 2021Fairness-Aware PAC Learning from Corrupted Datacited by 18
  34. 2021Leveraging Expert Consistency to Improve Algorithmic Decision Supportcited by 17
  35. 2020How Does a Neural Network's Architecture Impact Its Robustness to Noisy Labels?cited by 12
  36. 2021Logit-based Uncertainty Measure in Classificationcited by 3

Built on by

Cited by year

Citing papers per year: 2023: 1; 2024: 5; 2025: 1.

7 papers.

  1. 2024Usable XAI: 10 Strategies Towards Exploiting Explainability in the LLM Eracited by 45
  2. 2024Automating Data Annotation under Strategic Human Agents: Risks and Potential Solutionscited by 9
  3. 2024The Fragility of Fairness: Causal Sensitivity Analysis for Fair Machine Learningcited by 8
  4. 2025FNBench: Benchmarking Robust Federated Learning against Noisy Labelscited by 6
  5. 2024Data Debiasing with Datamodels (D3M): Improving Subgroup Robustness via Data Selectioncited by 5
  6. 2023Elucidating and Overcoming the Challenges of Label Noise in Supervised Contrastive Learningcited by 2
  7. 2024Wasserstein Distributionally Robust Multiclass Support Vector Machinecited by 0