A Review On Game Theoretical Incentive Mechanism For Federated Learning

Rapid growth of training data has led the Artificial Intelligence industry to the distributed learning paradigms. Federated learning has recently gained attention for the privacy preserved distributed learning activities. This collaborative training method facilitates training of data where it originates and secure aggregation of all the local models at the central server. In order to inspire the clients for the active and long run participation in the learning activity, appropriate incentive mechanisms has been developed. In this paper, a systematic review of incentive mechanisms for federated learning based on game theory concept has been studied.