Learning Controllable Fair Representations
This work addresses fairness in machine learning for reducing unfair decisions, but it is incremental as it builds on existing approaches by introducing a controllable trade-off.
The paper tackles the problem of learning data representations that are fair with respect to protected attributes while maintaining utility, proposing an information-theoretic objective that allows user control over fairness limits and achieves higher expressiveness at lower computational cost.
Learning data representations that are transferable and are fair with respect to certain protected attributes is crucial to reducing unfair decisions while preserving the utility of the data. We propose an information-theoretically motivated objective for learning maximally expressive representations subject to fairness constraints. We demonstrate that a range of existing approaches optimize approximations to the Lagrangian dual of our objective. In contrast to these existing approaches, our objective allows the user to control the fairness of the representations by specifying limits on unfairness. Exploiting duality, we introduce a method that optimizes the model parameters as well as the expressiveness-fairness trade-off. Empirical evidence suggests that our proposed method can balance the trade-off between multiple notions of fairness and achieves higher expressiveness at a lower computational cost.