Federated Learning using Smart Contracts on Blockchains, based on Reward\n Driven Approach

Abstract

Over the recent years, Federated machine learning continues to gain interest\nand momentum where there is a need to draw insights from data while preserving\nthe data provider's privacy. However, one among other existing challenges in\nthe adoption of federated learning has been the lack of fair, transparent and\nuniversally agreed incentivization schemes for rewarding the federated learning\ncontributors. Smart contracts on a blockchain network provide transparent,\nimmutable and independently verifiable proofs by all participants of the\nnetwork. We leverage this open and transparent nature of smart contracts on a\nblockchain to define incentivization rules for the contributors, which is based\non a novel scalar quantity - federated contribution. Such a smart contract\nbased reward-driven model has the potential to revolutionize the federated\nlearning adoption in enterprises. Our contribution is two-fold: first is to\nshow how smart contract based blockchain can be a very natural communication\nchannel for federated learning. Second, leveraging this infrastructure, we can\nshow how an intuitive measure of each agents' contribution can be built and\nintegrated with the life cycle of the training and reward process.\n