Fair Transfer Learning with Missing Protected Attributes

By Amanda Coston, Karthikeyan Natesan Ramamurthy, Dennis Wei, Kush R. Varshney, Skyler Speakman, Zairah Mustahsan, Supriyo Chakraborty. 2019

123 citations recorded by the source corpus.

Risk assessment is a growing use for machine learning models. When used in high-stakes applications, especially ones regulated by anti-discrimination laws or governed by societal norms for fairness, it is important to ensure that learned models do not propagate and scale any biases that may exist in training data. In this paper, we add on an additional challenge beyond fairness: unsupervised domain adaptation to covariate shift between a source and target distribution. Motivated by the real-world problem of risk assessment in new markets for health insurance in the United States and mobile money-based loans in East Africa, we provide a precise formulation of the machine learning with covariate shift and score parity problem. Our formulation focuses on situations in which protected attributes are not available in either the source or target domain. We propose two new weighting methods: prevalence-constrained covariate shift (PCCS) which does not require protected attributes in the target domain and target-fair covariate shift (TFCS) which does not require protected attributes in the source domain. We empirically demonstrate their efficacy in two applications.

Subfields: Artificial Intelligence, Safety Research

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Safety Research

Citations in the corpus

Characterizing Fairness Over the Set of Good Models Under Selective Labels (2021, methodology): “Reweighing procedures have been proposed for jointly addressing covariate shift and fairness (Coston et al., 2019; Singh et al., 2021).”
Fairness Violations and Mitigation under Covariate Shift (2019, background): “In the same setup as ours under the assumption of covariate shift but with the availability of unlabelled target data, [12] give weighting-based estimators and [59] take a robust optimization approach.”
Fairness without Demographics through Adversarially Reweighted Learning (2020, background): “Multiple works have explored addressing limited demographic data with transfer learning [11, 42, 47].”
A Survey on the Fairness of Recommender Systems (2022, background): “There has been some related work on classification tasks [17, 18, 48, 92, 100], while in recommender systems, it is still a problem to be explored.”
Fairness in Machine Learning: A Survey (2020, background): “…transfer learning approaches are somewhat analogous to reweighing approaches in that they seek to determine weights for each training example that account for a covariate shift optimized using regularization techniques (e.g., Reference [78]) or forms of joint loss functions (e.g., Reference [90]).”