Fairness without Demographics through Adversarially Reweighted Learning

Abstract

Much of the previous machine learning (ML) fairness literature assumes that protected features such as race and sex are present in the dataset, and relies upon them to mitigate fairness concerns. However, in practice factors like privacy and regulation often preclude the collection of protected features, or their use for training or inference, severely limiting the applicability of traditional fairness research. Therefore we ask: How can we train an ML model to improve fairness when we do not even know the protected group memberships? In this work we address this problem by proposing Adversarially Reweighted Learning (ARL). In particular, we hypothesize that non-protected features and task labels are valuable for identifying fairness issues, and can be used to co-train an adversarial reweighting approach for improving fairness. Our results show that {ARL} improves Rawlsian Max-Min fairness, with notable AUC improvements for worst-case protected groups in multiple datasets, outperforming state-of-the-art alternatives.

To appear at 34th Conference on Neural Information Processing Systems (NeurIPS 2020), Vancouver, Canada

References in corpus (16)

• Equality of Opportunity in Supervised Learning
• Preventing Fairness Gerrymandering: Auditing and Learning for Subgroup Fairness
• Data Decisions and Theoretical Implications when Adversarially Learning Fair Representations
• Agnostic Federated Learning
• Fairness Without Demographics in Repeated Loss Minimization
• From Parity to Preference-based Notions of Fairness in Classification
• Detecting and Correcting for Label Shift with Black Box Predictors
• Fairness Through Computationally-Bounded Awareness
• Operationalizing Individual Fairness with Pairwise Fair Representations
• Assessing Algorithmic Fairness with Unobserved Protected Class Using Data Combination
• An Empirical Study on Learning Fairness Metrics for COMPAS Data with Human Supervision
• Equalized odds postprocessing under imperfect group information
• Proxy Fairness
• Adversarial Classifier for Imbalanced Problems
• Environment Inference for Invariant Learning
• Toward a better trade-off between performance and fairness with kernel-based distribution matching

Cited by in corpus (5)

• A Comprehensive Survey on Trustworthy Graph Neural Networks: Privacy, Robustness, Fairness, and Explainability
• "What We Can't Measure, We Can't Understand": Challenges to Demographic Data Procurement in the Pursuit of Fairness
• Subgroup Fairness in Two-Sided Markets
• Tackling Racial Bias in Automated Online Hate Detection: Towards Fair and Accurate Classification of Hateful Online Users Using Geometric Deep Learning
• MABR: Multilayer Adversarial Bias Removal Without Prior Bias Knowledge