Average Individual Fairness: Algorithms, Generalization and Experiments

By Michael Kearns, Aaron Roth, Saeed Sharifi-Malvajerdi. 2019

98 citations recorded by the source corpus.

We propose a new family of fairness definitions for classification problems that combine some of the best properties of both statistical and individual notions of fairness. We posit not only a distribution over individuals, but also a distribution over (or collection of) classification tasks. We then ask that standard statistics (such as error or false positive/negative rates) be (approximately) equalized across individuals, where the rate is defined as an expectation over the classification tasks. Because we are no longer averaging over coarse groups (such as race or gender), this is a semantically meaningful individual-level constraint. Given a sample of individuals and classification problems, we design an oracle-efficient algorithm (i.e. one that is given access to any standard, fairness-free learning heuristic) for the fair empirical risk minimization task. We also show that given sufficiently many samples, the ERM solution generalizes in two directions: both to new individuals, and to new classification tasks, drawn from their corresponding distributions. Finally we implement our algorithm and empirically verify its effectiveness.

How researchers see this paper

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

Citations in the corpus

Controllable Guarantees for Fair Outcomes via Contrastive Information Estimation (2021, background): “Incorporating other popular notions of group fairness like Equalized-Opportunity and Equalized-Odds (Hardt et al. 2016), and notions of individual fairness (e.g., Shariﬁ-Malvajerdi, Kearns, and Roth (2019)) in our approach is left as future work.”
No Subclass Left Behind: Fine-Grained Robustness in Coarse-Grained Classification Problems (2020, background): “[26] Michael Kearns, Aaron Roth, and Saeed Sharifi-Malvajerdi.”