Proxy Fairness

By Maya R. Gupta, Andrew Cotter, Mahdi Milani Fard, Serena Wang. 2018

77 citations recorded by the source corpus.

We consider the problem of improving fairness when one lacks access to a dataset labeled with protected groups, making it difficult to take advantage of strategies that can improve fairness but require protected group labels, either at training or runtime. To address this, we investigate improving fairness metrics for proxy groups, and test whether doing so results in improved fairness for the true sensitive groups. Results on benchmark and real-world datasets demonstrate that such a proxy fairness strategy can work well in practice. However, we caution that the effectiveness likely depends on the choice of fairness metric, as well as how aligned the proxy groups are with the true protected groups in terms of the constrained model parameters.

Subfields: Artificial Intelligence, Safety Research, Sociology and Political Science

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

In Sociology and Political Science

Citations in the corpus

Beyond Individual and Group Fairness (2020, background): “A recent line of work has studied this estimation problem when one has limited and/or noisy access to sensitive attribute information [Gupta et al., 2018, Coston et al., 2019, Lamy et al., 2019, Wang et al., 2020].”
What’s in a Name? Reducing Bias in Bios without Access to Protected Attributes (2019, background): “Much like the “proxy fairness” approach of Gupta et al. (2018), it is applicable when protected attributes are not available; however, it additionally eliminates the need to specify which biases are to be mitigated, and allows simultaneous mitigation of multiple biases, including those that relate…”
Fairness for Unobserved Characteristics: Insights from Technological Impacts on Queer Communities (2021, background): “While it is possible to root the analysis in proxy groups (Gupta et al. 2018), there is a risk of incorporating harmful stereotypes in proxy group definitions, potentially resulting in harms of representation (Abbasi et al.”
What We Can't Measure, We Can't Understand: Challenges to Demographic Data Procurement in the Pursuit of Fairness (2020, neutral): “…proposed a number of methods and techniques that explicitly try to bypass the collection of demographic data through inference and proxies (e.g., [32, 69, 90]), to only use demographic data during model training (e.g., [28, 49, 71, 88]) 1 , or to identify and mitigate algorithmic discrimination…”
Fair Transfer Learning with Missing Protected Attributes (2019, background): “Gupta et al. [13] address the problem of unavailable protected attributes by constructing proxy groups using variables in the dataset that are not protected but are likely correlated to protected attributes based on prior subject matter knowledge.”
Fairness without Demographics through Adversarially Reweighted Learning (2020, background): “many real world applications the protected attribute information might be missing or is very sparse. Fairness without demographics: Some works address this issue approximately by using proxy features [14] or assuming that the attribute is slightly perturbed [3]. However, using proxies can in itself ”
Fairness in Machine Learning: A Survey (2020, background): “Reference [134] also learns proxy groupings.”