Beyond Distributive Fairness in Algorithmic Decision Making: Feature Selection for Procedurally Fair Learning

AAAI Conferences

With widespread use of machine learning methods in numerous domains involving humans, several studies have raised questions about the potential for unfairness towards certain individuals or groups. A number of recent works have proposed methods to measure and eliminate unfairness from machine learning models. However, most of this work has focused on only one dimension of fair decision making: distributive fairness, i.e., the fairness of the decision outcomes. In this work, we leverage the rich literature on organizational justice and focus on another dimension of fair decision making: procedural fairness, i.e., the fairness of the decision making process. We propose measures for procedural fairness that consider the input features used in the decision process, and evaluate the moral judgments of humans regarding the use of these features. We operationalize these measures on two real world datasets using human surveys on the Amazon Mechanical Turk (AMT) platform, demonstrating that our measures capture important properties of procedurally fair decision making. We provide fast submodular mechanisms to optimize the tradeoff between procedural fairness and prediction accuracy. On our datasets, we observe empirically that procedural fairness may be achieved with little cost to outcome fairness, but that some loss of accuracy is unavoidable.


Learning Individually Fair Classifier with Causal-Effect Constraint

arXiv.org Artificial Intelligence

Machine learning is increasingly being used to make decisions for individuals in lending, hiring, and recidivism prediction. Such applications require us to make decisions that are not discriminatory with respect to a sensitive feature, e.g., gender, race, or sexual orientation. This requirement is indispensable because data often exhibit a discriminatory bias including a correlation between the decision outcome and a sensitive feature if they include the records of past discriminatory decisions made by humans. Although many researchers have studied how to make fair decisions while achieving high prediction accuracy Feldman et al. [2015], Hardt et al. [2016], it remains a challenge in complex real-world scenarios. For instance, let us consider the case of making hiring decisions for applicants for physically demanding jobs (e.g., construction). While it is discriminatory to reject applicants because of their gender, since the job requires physical strength, it is sometimes not discriminatory to reject them because of their physical strength. Since physical strength is affected by gender, rejecting applicants because of physical strength leads to a difference in the rejection rates for men and women.


Grgić-Hlača

AAAI Conferences

With widespread use of machine learning methods in numerous domains involving humans, several studies have raised questions about the potential for unfairness towards certain individuals or groups. A number of recent works have proposed methods to measure and eliminate unfairness from machine learning models. However, most of this work has focused on only one dimension of fair decision making: distributive fairness, i.e., the fairness of the decision outcomes. In this work, we leverage the rich literature on organizational justice and focus on another dimension of fair decision making: procedural fairness, i.e., the fairness of the decision making process. We propose measures for procedural fairness that consider the input features used in the decision process, and evaluate the moral judgments of humans regarding the use of these features. We operationalize these measures on two real world datasets using human surveys on the Amazon Mechanical Turk (AMT) platform, demonstrating that our measures capture important properties of procedurally fair decision making. We provide fast submodular mechanisms to optimize the tradeoff between procedural fairness and prediction accuracy. On our datasets, we observe empirically that procedural fairness may be achieved with little cost to outcome fairness, but that some loss of accuracy is unavoidable.


Proxy Fairness

arXiv.org Machine Learning

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.


Differentially Private Fair Learning

arXiv.org Machine Learning

Large-scale algorithmic decision making, often driven by machine learning on consumer data, has increasingly run afoul of various social norms, laws and regulations. A prominent concern is when a learned model exhibits discrimination against some demographic group, perhaps based on race or gender. Concerns over such algorithmic discrimination have led to a recent flurry of research on fairness in machine learning, which includes both new tools and methods for designing fair models, and studies of the tradeoffs between predictive accuracy and fairness [ACM, 2019]. At the same time, both recent and longstanding laws and regulations often restrict the use of "sensitive" or protected attributes in algorithmic decision-making. U.S. law prevents the use of race in the development or deployment of consumer lending or credit scoring models, and recent provisions in the E.U. General Data Protection Regulation (GDPR) restrict or prevent even the collection of racial data for consumers. These two developments -- the demand for non-discriminatory algorithms and models on the one hand, and the restriction on the collection or use of protected attributes on the other -- present technical conundrums, since the most straightforward methods for ensuring fairness generally require knowing or using the attribute being protected. It seems difficult to guarantee that a trained model is not discriminating against, say, a racial group if we cannot even identify members of that group in the data. 1 A recent paper [Kilbertus et al., 2018] made these cogent observations, and proposed an interesting solutionemploying the cryptographic tool of secure multiparty computation (commonly abbreviated MPC). In their model, we imagine a commercial entity with access to consumer data that excludes race, but this entity would like to build a predictive model for, say, commercial lending, under the constraint that the model be non-discriminatory by race with respect to some standard fairness notion (e.g.