Blockchain-Based Federated Learning for Incentivizing Data Sharing & Penalizing Dishonest Behavior
Blockchain-Based Federated Learning: Incentivizing Data Sharing and Penalizing Dishonest Behavior
Amir Jaberzadeh, Ajay Kumar Shrestha, Faijan Ahamad Khan, Mohammed Afaan Shaikh, Bhargav Dave, and Jason Geng
Bayes Solutions, 840 Apollo St, El Segundo CA 90245, USA
Vancouver Island University, 900 Fifth St, Nanaimo, BC V9R 5S5, Canada
arXiv: 2307.10492
Federated Learning is a novel machine learning paradigm in which a model is trained among distributed participants on local data. Aggregating the individual models with a central server or using decentralized techniques results in a final model that profits from all the local data of the user without having to share it. In this work, we present an architecture for decentralized Federated Learning that uses blockchain to distribute rewards among the participants. We define the notion of trust in such a system and show how our architecture implements trustworthiness. To support our claims, we deliver a prototype application that allows simulating the full architecture.
For more information, read the full paper.
In the client folder, you can find the code and instructions to execute it.