Incentive-Boosted Federated Crowdsourcing

Crowdsourcing is a favorable computing paradigm for processing computer-hard tasks by harnessing human intelligence. However, generic crowdsourcing systems may lead to privacy-leakage through the sharing of worker data. To tackle this problem, we propose novel approach, called iFedCrowd (incentive-boosted Federated Crowdsourcing), manage privacy and quality projects. allows participants locally process sensitive data only upload encrypted training models, then aggregates model parameters build shared server protect privacy. motivate workers high-quality global in an efficacy way, introduce incentive mechanism that encourages constantly collect fresh train accurate client models boosts training. We incentive-based interaction between platform participating as Stackelberg game, which each side maximizes its own profit. derive Nash Equilibrium game find optimal solutions two sides. Experimental results confirm can complete secure projects with high efficiency.