Fairness-oriented vertical federated GNNs with incomplete sensitive attributes

doi:10.1016/j.inffus.2025.103871

Fairness-oriented vertical federated GNNs with incomplete sensitive attributes

2025-10-19

0

PRE

AI

DOI：10.1016/j.inffus.2025.103871

Original

Original request for help

Save

Abstract

En 中文

• Proposes FaVGNN, the first fairness-oriented framework in vertical FL. • Uses completion-driven adversarial fusion to infer missing attributes and ensure fairness. • Introduces weighted aggregation to balance accuracy and fairness across clients. • Experiments show FaVGNN outperforms baselines in fairness and privacy preservation.

Fairness-oriented vertical federated GNNs with incomplete sensitive attributes

Abstract

Journal

Organization