Fair Representations by Compression
This addresses fairness concerns in data collection and sale for organizations, though it is incremental as it builds on information bottleneck frameworks.
The paper tackles the problem of discriminatory data use by proposing an unsupervised method to compress data into binary representations independent of sensitive attributes, achieving state-of-the-art accuracy-fairness trade-offs with explicit control over entropy.
Organizations that collect and sell data face increasing scrutiny for the discriminatory use of data. We propose a novel unsupervised approach to transform data into a compressed binary representation independent of sensitive attributes. We show that in an information bottleneck framework, a parsimonious representation should filter out information related to sensitive attributes if they are provided directly to the decoder. Empirical results show that the proposed method, \textbf{FBC}, achieves state-of-the-art accuracy-fairness trade-off. Explicit control of the entropy of the representation bit stream allows the user to move smoothly and simultaneously along both rate-distortion and rate-fairness curves. \end{abstract}