Improving Privacy-Preserving Vertical Federated Learning by Efficient Communication with ADMM

arXiv:2207.10226v4 Announce Type: replace-cross
Abstract: Federated learning (FL) enables distributed resource-constrained devices to jointly train shared models while keeping the training data local for privacy purposes. Vertical FL (VFL), which allows each client to collect partial features, has attracted intensive research efforts recently. We identified the main challenges that existing VFL frameworks are facing: the server needs to communicate gradients with the clients for each training step, incurring high communication cost that leads to rapid consumption of privacy budgets. To …

arxiv challenges client collect communication cs.cr cs.lg data devices distributed features federated federated learning local main partial privacy research resource train training training data