The Variational Fair Autoencoder
This addresses fairness and bias reduction in machine learning models by purging sensitive information from representations, though it appears incremental as it builds on existing variational autoencoder techniques.
The paper tackles the problem of learning data representations that are invariant to sensitive factors while preserving other information, using a variational autoencoder with priors and an MMD penalty, and shows it is more effective than previous methods.
We investigate the problem of learning representations that are invariant to certain nuisance or sensitive factors of variation in the data while retaining as much of the remaining information as possible. Our model is based on a variational autoencoding architecture with priors that encourage independence between sensitive and latent factors of variation. Any subsequent processing, such as classification, can then be performed on this purged latent representation. To remove any remaining dependencies we incorporate an additional penalty term based on the "Maximum Mean Discrepancy" (MMD) measure. We discuss how these architectures can be efficiently trained on data and show in experiments that this method is more effective than previous work in removing unwanted sources of variation while maintaining informative latent representations.