Two-Layered Blockchain Architecture for Federated Learning Over the Mobile Edge Network

Federated learning (FL) is seen as a road toward privacy-preserving distributed artificial intelligence while keeping raw training data on local devices. By leveraging blockchain, this article puts forward a blockchain and FL fusioned framework to manage the security and trust issues when applying FL over mobile edge networks. First, a two-layered architecture is proposed that consists of two types of blockchains: local model update chain (LMUC) assisted by device-to-device (D2D) communication and global model update chain (GMUC) supporting task sharding. The D2D-assisted LMUC is designed to chronologically and efficiently record all of the local model training results, which can help to form long-term reputations of local devices. The GMUC is proposed to provide both security and efficiency by preventing mobile edge computing nodes from malfunctioning and dividing them into logically isolated FL task-specific chains. Then a reputation-learning-based incentive mechanism is introduced to make participating local devices more trustful with a reward implemented by a smart contract. Finally, a case study is given to show that the proposed framework performs well in terms of FL learning accuracy and blockchain time delay.