Demystifying Local and Global Fairness Trade-offs in Federated Learning Using Partial Information Decomposition

arXiv:2307.11333v2 Announce Type: replace
Abstract: This work presents an information-theoretic perspective to group fairness trade-offs in federated learning (FL) with respect to sensitive attributes, such as gender, race, etc. Existing works often focus on either $\textit{global fairness}$ (overall disparity of the model across all clients) or $\textit{local fairness}$ (disparity of the model at each client), without always considering their trade-offs. There is a lack of understanding regarding the interplay between global and local fairness in FL, particularly under data heterogeneity, …

abstract arxiv cs.cy cs.it cs.lg etc fairness federated learning focus gender global information math.it perspective race trade type work