| Title |
Blockchain-based Federated Learning Reference Architecture and DID Access System |
| Authors |
고은수(Eunsu Goh) ; 김대열(Daeyeol Kim) ; 이광기(Kwangkee Lee) |
| DOI |
https://doi.org/10.5573/ieie.2023.60.7.50 |
| Keywords |
Federated learning; Blockchain; Reference architecture; Deep learning; Machine learning |
| Abstract |
The blockchain-based federated learning architecture is a novel approach that combines the advantages of federated learning and blockchain technology. Through this architecture, multiple participants can train machine learning models in a distributed manner while maintaining data privacy and security. To ensure secure access in this architecture, a Decentralized Identifier (DID) based access system can be used. DID provides a decentralized and self-sovereign ID management system that allows users to own and control their IDs without relying on central authorities. In this architecture, participants can authenticate themselves and access the federated learning platform using DIDs. DIDs are stored on the blockchain, and the access system manages access control and permissions using smart contracts. The combination of the blockchain-based federated learning architecture and the DID-based access system provides a secure and distributed approach to collaborative machine learning in a decentralized and secure environment. Participants can contribute to global model training while maintaining data privacy and ID control without sharing local data. This approach offers a promising solution for collaborative machine learning in a distributed and secure environment and can be effectively integrated with the design of blockchain-based federated learning architectures. |