Fairness risk measures

Abstract

Ensuring that classifiers are non-discriminatory or fair with respect to a sensitive feature (e.g., race or gender) is a topical problem. Progress in this task requires fixing a definition of fairness, and there have been several proposals in this regard over the past few years. Several of these, however, assume either binary sensitive features (thus precluding categorical or real-valued sensitive groups), or result in non-convex objectives (thus adversely affecting the optimisation landscape). In this paper, we propose a new definition of fairness that generalises some existing proposals, while allowing for generic sensitive features and resulting in a convex objective. The key idea is to enforce that the expected losses (or risks) across each subgroup induced by the sensitive feature are commensurate. We show how this relates to the rich literature on risk measures from mathematical finance. As a special case, this leads to a new convex fairness-aware objective based on minimising the conditional value at risk (CVaR).

References in corpus (8)

• Equality of Opportunity in Supervised Learning
• Counterfactual Fairness
• Empirical Bernstein Bounds and Sample Variance Penalization
• From Parity to Preference-based Notions of Fairness in Classification
• Learning with Average Top-k Loss
• On conditional parity as a notion of non-discrimination in machine learning
• Two-stage Algorithm for Fairness-aware Machine Learning
• Unleashing Linear Optimizers for Group-Fair Learning and Optimization

Cited by in corpus (13)

• Out-of-Distribution Generalization via Risk Extrapolation (REx)
• Noise-tolerant fair classification
• Learning Bounds for Risk-sensitive Learning
• PAC-Bayesian Bound for the Conditional Value at Risk
• Adaptive Sampling for Stochastic Risk-Averse Learning
• Environment Inference for Invariant Learning
• Optimal Best-Arm Identification Methods for Tail-Risk Measures
• Coping with Label Shift via Distributionally Robust Optimisation
• Noisy Linear Convergence of Stochastic Gradient Descent for CV@R Statistical Learning under Polyak-Łojasiewicz Conditions
• Fairness Through Regularization for Learning to Rank
• A Survey of Learning Criteria Going Beyond the Usual Risk
• Differential Privacy, Linguistic Fairness, and Training Data Influence: Impossibility and Possibility Theorems for Multilingual Language Models
• Exploring the Unfairness of DP-SGD Across Settings