Proof of Federated Learning Contribution (PoFLC) for Secure Agri-Data Sharing and Collaborative Model Training

Abstract

The proliferation of heterogeneous agricultural devices generates vast amounts of data, necessitating secure and efficient methods for decentralized training and data sharing. Traditional centralized approaches face significant challenges concerning data privacy, security, and scalability. This paper introduces a novel consensus mechanism called Proof of Federated Learning Contribution (PoFLC) within a secure multi-chain framework designed for agri-data sharing and collaborative model training. PoFLC quantifies the contributions of participating devices based on model update quality, reliability, and resources. Integrating federated learning with blockchain technology ensures data privacy while enhancing model performance in a decentralized environment. Extensive experiments on agricultural datasets demonstrate the effectiveness of PoFLC in maintaining data confidentiality, achieving consensus efficiently, and improving the accuracy of collaborative models trained on agri-data.