FairGAN: GANs-based Fairness-aware Learning for Recommendations with Implicit Feedback

In this paper, we introduce the Fairness GAN, an approach for generating a dataset that is plausibly similar to a given multimedia dataset, but is more fair with respect to protected attributes in allocative decision making. We propose a novel auxiliary classifier GAN that strives for demographic parity or equality of opportunity and show empirical results on several datasets, including the CelebFaces Attributes (CelebA) dataset, the Quick, Draw! dataset, and a dataset of soccer player images and the offenses they were called for. The proposed formulation is well-suited to absorbing unlabeled data; we leverage this to augment the soccer dataset with the much larger CelebA dataset. The methodology tends to improve demographic parity and equality of opportunity while generating plausible images. Recent contributions examining fairness from the perspective of adversarial learning include references [22] [23] [24] [25] . These methods are similar to our proposed approach by including a classifier trained to perform as poorly as possible on predicting the outcome from the protected attribute, but are different from our proposed method in two key ways. First, they are not intended to create Preprint. Work in progress.