Abstract

Blockchain is one of the most secure, reliable, decentralized, and distributed technologies that has emerged today, which forms the backbone of many applications in modern times across sectors ranging from banking to finance, insurance, healthcare, and business. Its capability for transparency, immutability, and tamper resistance has attracted various types of stakeholders who increasingly look to improve on-chain processes with intelligent data-driven capabilities. This paper is important because, in recent years, a growing volume of geographically dispersed, privately-owned data within blockchain networks has motivated an interest in incorporating machine learning models to derive meaningful insights and enhance decision-making. However, traditional machine learning relies heavily on centralized data collection, contradicting the decentralized nature of blockchain and putting it at serious privacy risk.