Efficient fair PCA for fair representation learning
This addresses fairness in representation learning for applications requiring demographic privacy, though it is incremental as it builds on existing fair PCA approaches.
The authors tackled the problem of fair principal component analysis (PCA) by developing a simple method that learns low-rank linear approximations while obfuscating demographic information, achieving similar results to existing methods but with much faster runtime comparable to standard PCA.
We revisit the problem of fair principal component analysis (PCA), where the goal is to learn the best low-rank linear approximation of the data that obfuscates demographic information. We propose a conceptually simple approach that allows for an analytic solution similar to standard PCA and can be kernelized. Our methods have the same complexity as standard PCA, or kernel PCA, and run much faster than existing methods for fair PCA based on semidefinite programming or manifold optimization, while achieving similar results.