An Optimized and Scalable Blockchain-Based Distributed Learning Platform for Consumer IoT

Abstract

Consumer Internet of Things (CIoT) manufacturers seek customer feedback to enhance their products and services, creating a smart ecosystem, like a smart home. Due to security and privacy concerns, blockchain-based federated learning (BCFL) ecosystems can let CIoT manufacturers update their machine learning (ML) models using end-user data. Federated learning (FL) uses privacy-preserving ML techniques to forecast customers’ needs and consumption habits, and blockchain replaces the centralized aggregator to safeguard the ecosystem. However, blockchain technology (BCT) struggles with scalability and quick ledger expansion. In BCFL, local model generation and secure aggregation are other issues. This research introduces a novel architecture, emphasizing gateway peer (GWP) in the blockchain network to address scalability, ledger optimization, and secure model transmission issues. In the architecture, we replace the centralized aggregator with the blockchain network, while GWP limits the number of local transactions to execute in BCN. Considering the security and privacy of FL processes, we incorporated differential privacy and advanced normalization techniques into ML processes. These approaches enhance the cybersecurity of end-users and promote the adoption of technological innovation standards by service providers. The proposed approach has undergone extensive testing using the well-respected Stanford (CARS) dataset. We experimentally demonstrate that the proposed architecture enhances network scalability and significantly optimizes the ledger. In addition, the normalization technique outperforms batch normalization when features are under DP protection.

Keywords

• blockchain
• IoT
• security and privacy
• smart home
• distributed ledger technology