×
Nov 3, 2022 · We model the task of learning a shared predictor in the federated setting as a fair public decision making problem, and then define the notion of core-stable ...
Federated learning provides an effective paradigm to jointly optimize a model benefited from rich distributed data while protecting data privacy.
Apr 3, 2024 · In this work, we aim to formally represent this problem and address these fairness issues using concepts from co-operative game theory and ...
We design an effective FL protocol CoreFed to realize core-stable training when possible. On three datasets, CoreFed achieves core-stable fairness, while.
u(θ) = M − E(x,y)∼P ℓ (1) where M is a constant more than (1 + ε) times the loss incurred from the worst predictor for agent.
Nov 3, 2022 · The fairness of federated learning means that the global model cannot discriminate against any group. Due data heterogeneity, the update ...
Jul 29, 2024 · Previous work on fairness in federated learning introduced the notion of core stability, which provides utility-based fairness guarantees to ...
Jan 1, 2022 · Award ID(s):: 1750436 ; NSF-PAR ID: 10403774 ; Author(s) / Creator(s):: Chaudhury, Bhaskar Ray; Li, Linyi; Kang, Mintong; Li, Bo; Mehta, Ruta.
Previous work on fairness in federated learning introduced the notion of core stability, which pro- vides utility-based fairness guarantees to any sub-.
People also ask
Fairness in Federated Learning via Core-Stability. Bhaskar Ray Chaudhury, Linyi Li, Mintong Kang, Bo Li, Ruta Mehta. Proceedings of the 36th Conference on ...