On Disentangled and Locally Fair Representations
This work addresses fairness in machine learning for sensitive groups, offering a novel approach to mitigate bias in classification tasks.
The paper tackles the problem of fair classification for sensitive groups like race and gender by learning disentangled and locally fair representations, ensuring that neighborhoods of samples are balanced in terms of sensitive attributes without using correlated features, and demonstrates its advantage in real-world settings such as income and re-incarceration rate prediction.
We study the problem of performing classification in a manner that is fair for sensitive groups, such as race and gender. This problem is tackled through the lens of disentangled and locally fair representations. We learn a locally fair representation, such that, under the learned representation, the neighborhood of each sample is balanced in terms of the sensitive attribute. For instance, when a decision is made to hire an individual, we ensure that the $K$ most similar hired individuals are racially balanced. Crucially, we ensure that similar individuals are found based on attributes not correlated to their race. To this end, we disentangle the embedding space into two representations. The first of which is correlated with the sensitive attribute while the second is not. We apply our local fairness objective only to the second, uncorrelated, representation. Through a set of experiments, we demonstrate the necessity of both disentangled and local fairness for obtaining fair and accurate representations. We evaluate our method on real-world settings such as predicting income and re-incarceration rate and demonstrate the advantage of our method.