mlsanityData to 14 Oct 2026

On the Fairness of Disentangled Representations

arXiv 1905.13662 ↗First posted 31 May 2019cs.LG stat.MLcited by 144

Not on the map. Why?

Abstract (arXiv)

Recently there has been a significant interest in learning disentangled representations, as they promise increased interpretability, generalization to unseen scenarios and faster learning on downstream tasks. In this paper, we investigate the usefulness of different notions of disentanglement for improving the fairness of downstream prediction tasks based on representations. We consider the setting where the goal is to predict a target variable based on the learned representation of high-dimensional observations (such as images) that depend on both the target variable and an unobserved sensitive variable. We show that in this setting both the optimal and empirical predictions can be unfair, even if the target variable and the sensitive variable are independent. Analyzing the representations of more than num12600 trained state-of-the-art disentangled models, we observe that several disentanglement scores are consistently correlated with increased fairness, suggesting that disentanglement may be a useful property to encourage fairness when sensitive variables are not observed.

Key facts. Cited by 144 papers in the map (data through 14 Oct 2026). First posted on arXiv 31 May 2019.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2021: 41; 2022: 34; 2023: 29; 2024: 20; 2025: 13; 2026: 7.

144 papers.

Sort:

By area:

  1. 2021Towards Causal Representation Learningcited by 209
  2. 2022A Comprehensive Survey on Trustworthy Graph Neural Networks: Privacy, Robustness, Fairness, and Explainabilitycited by 118
  3. 2021Generalization and Robustness Implications in Object-Centric Learningcited by 81
  4. 2022Fair Contrastive Learning for Facial Attribute Classificationcited by 71
  5. 2024Preserving Fairness Generalization in Deepfake Detectioncited by 70
  6. 2022MEDFAIR: Benchmarking Fairness for Medical Imagingcited by 68
  7. 2023Compositional Generalization from First Principlescited by 64
  8. 2021Identifiable Deep Generative Models via Sparse Decodingcited by 54
  9. 2022Assaying Out-Of-Distribution Generalization in Transfer Learningcited by 53
  10. 2021Discover the Unknown Biased Attribute of an Image Classifiercited by 46
  11. 2021Learning Disentangled Representations in the Imaging Domaincited by 46
  12. 2021Towards Fair Classifiers Without Sensitive Attributes: Exploring Biases in Related Featurescited by 45
38 more
  1. 2021Fair Normalizing Flowscited by 38
  2. 2022Disentanglement of Correlated Factors via Hausdorff Factorized Supportcited by 37
  3. 2022Invariant Causal Mechanisms through Distribution Matchingcited by 31
  4. 2021An Image is Worth More Than a Thousand Words: Towards Disentanglement in the Wildcited by 30
  5. 2021Where and What? Examining Interpretable Disentangled Representationscited by 27
  6. 2023Survey of Social Bias in Vision-Language Modelscited by 26
  7. 2021A Sociotechnical View of Algorithmic Fairnesscited by 24
  8. 2022Bayes-Optimal Classifiers under Group Fairnesscited by 24
  9. 2021Towards Building A Group-based Unsupervised Representation Disentanglement Frameworkcited by 22
  10. 2022Fair Representation Learning through Implicit Path Alignmentcited by 22
  11. 2022Debiasing Methods for Fairer Neural Models in Vision and Language Research: A Surveycited by 22
  12. 2023From Identifiable Causal Representations to Controllable Counterfactual Generation: A Survey on Causal Generative Modelingcited by 22
  13. 2021Algorithm Fairness in AI for Medicine and Healthcarecited by 21
  14. 2021EditVAE: Unsupervised Part-Aware Controllable 3D Point Cloud Shape Generationcited by 21
  15. 2021A Near-Optimal Algorithm for Debiasing Trained Machine Learning Modelscited by 19
  16. 2021Drop, Swap, and Generate: A Self-Supervised Approach for Generating Neural Activitycited by 18
  17. 2021Latent Space Smoothing for Individually Fair Representationscited by 18
  18. 2021Modeling Techniques for Machine Learning Fairness: A Surveycited by 17
  19. 2022Representational dissimilarity metric spaces for stochastic neural networkscited by 17
  20. 2021Disentangling Representations of Text by Masking Transformerscited by 15
  21. 2021Addressing the Topological Defects of Disentanglement via Distributed Operatorscited by 14
  22. 2023Identifiable Contrastive Learning with Automatic Feature Importance Discoverycited by 14
  23. 2021Disentangled Representations from Non-Disentangled Modelscited by 13
  24. 2023Towards a Unified Framework of Contrastive Learning for Disentangled Representationscited by 13
  25. 2024Fair Graph Representation Learning via Sensitive Attribute Disentanglementcited by 13
  26. 2021The Role of Pretrained Representations for the OOD Generalization of Reinforcement Learning Agentscited by 12
  27. 2021DISP6D: Disentangled Implicit Shape and Pose Learning for Scalable 6D Pose Estimationcited by 12
  28. 2021On Characterizing the Trade-off in Invariant Representation Learningcited by 12
  29. 2024Complexity Matters: Dynamics of Feature Learning in the Presence of Spurious Correlationscited by 12
  30. 2022Is Fairness Only Metric Deep? Evaluating and Addressing Subgroup Gaps in Deep Metric Learningcited by 11
  31. 2022Certifying Some Distributional Fairness with Subpopulation Decompositioncited by 11
  32. 2023Fair Off-Policy Learning from Observational Datacited by 11
  33. 2023Measuring Bias in AI Models: An Statistical Approach Introducing N-Sigmacited by 11
  34. 2021Exploring the Latent Space of Autoencoders with Interventional Assayscited by 10
  35. 2022How do Variational Autoencoders Learn? Insights from Representational Similaritycited by 10
  36. 2023Simple Disentanglement of Style and Content in Visual Representationscited by 10
  37. 2021On Fair Selection in the Presence of Implicit and Differential Variancecited by 9
  38. 2022Learning Disentangled Behaviour Patterns for Wearable-based Human Activity Recognitioncited by 9

50 most cited of 144 All 144 (JSON)