mlsanityData to 14 Oct 2026

No Fair Lunch: A Causal Perspective on Dataset Bias in Machine Learning for Medical Imaging

arXiv 2307.16526 ↗First posted 31 Jul 2023cs.LG cs.AI cs.CVcited by 11

Not on the map. Why?

Abstract (arXiv)

As machine learning methods gain prominence within clinical decision-making, addressing fairness concerns becomes increasingly urgent. Despite considerable work dedicated to detecting and ameliorating algorithmic bias, today's methods are deficient with potentially harmful consequences. Our causal perspective sheds new light on algorithmic bias, highlighting how different sources of dataset bias may appear indistinguishable yet require substantially different mitigation strategies. We introduce three families of causal bias mechanisms stemming from disparities in prevalence, presentation, and annotation. Our causal analysis underscores how current mitigation methods tackle only a narrow and often unrealistic subset of scenarios. We provide a practical three-step framework for reasoning about fairness in medical imaging, supporting the development of safe and equitable AI prediction models.

Key facts. Cited by 11 papers in the map (data through 14 Oct 2026). Builds on 38 of its 95 references that are in the map. First posted on arXiv 31 Jul 2023.

Builds on

38 of 95 references.

References · 38

  1. 2016Equality of Opportunity in Supervised Learningcited by 2,007
  2. 2019CheXpert: A Large Chest Radiograph Dataset with Uncertainty Labels and Expert Comparisoncited by 1,423
  3. 2018Datasheets for Datasetscited by 1,386
  4. 2018Model Cards for Model Reportingcited by 1,080
  5. 2018Learning Adversarially Fair and Transferable Representationscited by 353
  6. 2018Learning Not to Learn: Training Deep Neural Networks with Biased Datacited by 255
  7. 2019Towards Fairness in Visual Recognition: Effective Strategies for Bias Mitigationcited by 252
  8. 2018Women also Snowboard: Overcoming Bias in Captioning Modelscited by 236
  9. 2019Hidden Stratification Causes Clinically Meaningful Failures in Machine Learning for Medical Imagingcited by 216
  10. 2020Deep Structural Causal Models for Tractable Counterfactual Inferencecited by 211
  11. 2018A comparative study of fairness-enhancing interventions in machine learningcited by 208
  12. 2019Causality matters in medical imagingcited by 193
26 more
  1. 2015Censoring Representations with an Adversarycited by 192
  2. 2018Path-Specific Counterfactual Fairnesscited by 181
  3. 2022Data Cards: Purposeful and Transparent Dataset Documentation for Responsible AIcited by 171
  4. 2019REPAIR: Removing Representation Bias by Dataset Resamplingcited by 157
  5. 2021Reading Race: AI Recognises Patient's Racial Identity In Medical Imagescited by 156
  6. 2020CheXclusion: Fairness gaps in deep chest X-ray classifierscited by 153
  7. 2019Identifying and Correcting Label Bias in Machine Learningcited by 146
  8. 2017Domain Adaptation by Using Causal Inference to Predict Invariant Conditional Distributionscited by 126
  9. 2020Fair Attribute Classification through Latent Space De-biasingcited by 125
  10. 2018Turning a Blind Eye: Explicit Removal of Biases and Variation from Deep Neural Network Embeddingscited by 114
  11. 2021DECAF: Generating Fair Synthetic Data Using Causally-Aware Generative Networkscited by 109
  12. 2018The Measure and Mismeasure of Fairnesscited by 95
  13. 2022MEDFAIR: Benchmarking Fairness for Medical Imagingcited by 68
  14. 2013Causal Networks: Semantics and Expressivenesscited by 67
  15. 2022Leveling Down in Computer Vision: Pareto Inefficiencies in Fair Deep Classifierscited by 53
  16. 2023High Fidelity Image Counterfactuals with Probabilistic Causal Modelscited by 53
  17. 2022Diagnosing failures of fairness transfer across distribution shift in real-world medical settingscited by 45
  18. 2022Causal Fairness Analysiscited by 43
  19. 2021Transporting Causal Mechanisms for Unsupervised Domain Adaptationcited by 38
  20. 2022Invariant and Transportable Representations for Anti-Causal Domain Shiftscited by 37
  21. 2018Active Fairness in Algorithmic Decision Makingcited by 27
  22. 2022Causal Transportability for Visual Recognitioncited by 23
  23. 2022Potential sources of dataset bias complicate investigation of underdiagnosis by machine learning algorithmscited by 19
  24. 2023The Role of Subgroup Separability in Group-Fair Medical Image Classificationcited by 11
  25. 2013Identifying Independencies in Causal Graphs with Feedbackcited by 7
  26. 2018Preventing Failures Due to Dataset Shift: Learning Predictive Models That Transportcited by 6

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2023: 2; 2024: 4; 2025: 2; 2026: 3.

11 papers.

  1. 2023A Multi-Center Study on the Adaptability of a Shared Foundation Model for Electronic Health Recordscited by 31
  2. 2023RadEdit: stress-testing biomedical vision models via diffusion image editingcited by 28
  3. 2024Abstraction Alignment: Comparing Model-Learned and Human-Encoded Conceptual Relationshipscited by 7
  4. 2024Semi-Supervised Learning for Deep Causal Generative Modelscited by 6
  5. 2024Trustworthy and Responsible AI for Human-Centric Autonomous Decision-Making Systemscited by 3
  6. 2024Causality-Driven Audits of Model Robustnesscited by 2
  7. 2025CAPRI-CT: Causal Analysis and Predictive Reasoning for Image Quality Optimization in Computed Tomographycited by 1
  8. 2025DISCO: Mitigating Bias in Deep Learning with Conditional Distance Correlationcited by 0
  9. 2026Causal Transfer in Medical Image Analysiscited by 0
  10. 2026Counterfactual Marginalisation: Framework for Evaluating Robustness to Nuisance Variablescited by 0
  11. 2026To do(x) or not to do(x): Medical Image Counterfactuals for Dataset Augmentationcited by 0