[KDD2021] Federated Adversarial Debiasing for Fair and Transferable Representations: Optimize an adversarial domain-adaptation objective without adversarial or source data. - View it on GitHub