A Fair Incentive Mechanism in Federated Learning

Abstract

Federated learning (FL) has shown great potential for addressing the challenge of isolated data islands while preserving data privacy. However, not everyone is willing to participate in federated learning to share their data computation results. Providing incentives is a common way to promote the success of federated learning. The design of this paper focuses on the incentive mechanism based on the reverse auction. First, an incentive model for crowdsourcing systems is constructed. Secondly, an auction-based incentive mechanism is proposed by combining the concepts of reverse auction and VCG auction. Simulation results indicate that the proposed incentive mechanism can effectively improve the fairness of the bids. In this paper, we present a VCG-based FL incentive mechanism, named LVCG, specifically designed for incentivizing data owners to contribute all their data and truthfully report their costs in FL settings.