Execution of a Federated Learning process within a smart contract

Abstract

High quality datasets have always been valuable for the creation of Machine Learning (ML) models. It therefore makes sense to provide rewards to users that participate in a Federated Learning (FL) process with such datasets. In this competitive scene, we design a solution that leverages a blockchain network, a smart contract and a model verification algorithm in order to coordinate the training process, record user performance and provide rewards in a transparent manner.