Learning Fair and Transferable Representations with Theoretical Guarantees | IEEE Conference Publication | IEEE Xplore