Optimal Fair Learning Robust to Adversarial Distribution Shift | Read Paper on Bytez