GIFAIR-FL: A Framework for Group and Individual Fairness in Federated Learning

In this paper, we propose GIFAIR-FL, a framework that imposes group and individual fairness (GIFAIR) to federated learning (FL) settings. By adding regularization term, our algorithm penalizes the spread in loss of client groups drive optimizer fair solutions. Our GIFAIR-FL can accommodate both global personalized Theoretically, show convergence nonconvex strongly convex guarantees hold for independent identically distributed (i.i.d.) non-i.i.d. data. To demonstrate empirical performance algorithm, apply method image classification text prediction tasks. Compared with existing algorithms, shows improved results while retaining superior or similar accuracy.