Are Two Heads the Same as One? Identifying Disparate Treatment in Fair Neural Networks
This addresses fairness in machine learning for applications like hiring or lending, but it is incremental as it builds on existing demographic parity methods.
The paper tackles the problem of demographic parity in fair neural networks by showing that enforcing fairness often makes networks more aware of protected attributes like race or gender, and proposes a two-headed network approach to explicitly predict these attributes and merge heads to achieve fairness. The result demonstrates that existing fairness methods, including theirs, exhibit disparate treatment and may be unlawful under US law.
We show that deep networks trained to satisfy demographic parity often do so through a form of race or gender awareness, and that the more we force a network to be fair, the more accurately we can recover race or gender from the internal state of the network. Based on this observation, we investigate an alternative fairness approach: we add a second classification head to the network to explicitly predict the protected attribute (such as race or gender) alongside the original task. After training the two-headed network, we enforce demographic parity by merging the two heads, creating a network with the same architecture as the original network. We establish a close relationship between existing approaches and our approach by showing (1) that the decisions of a fair classifier are well-approximated by our approach, and (2) that an unfair and optimally accurate classifier can be recovered from a fair classifier and our second head predicting the protected attribute. We use our explicit formulation to argue that the existing fairness approaches, just as ours, demonstrate disparate treatment and that they are likely to be unlawful in a wide range of scenarios under US law.