【Author】
Zhang, Xu; Hou, Haibo; Fang, Zhao; Wang, Zhiqian
【Source】WIRELESS COMMUNICATIONS & MOBILE COMPUTING
【Abstract】With the development of Internet of Things (IoT), 5G, and industrial technology, Industrial Internet has become an emerging research field. Due to the industrial specialty, higher requirements are put forward for time delay, safety, and stability of the identification analysis service. The traditional domain name system (DNS) cannot meet the requirements of industrial Internet because of the single form of identification subject and weak awareness of security protection. As a solution, this work applies blockchain and federated learning (FL) to the industrial Internet identification. Blockchain is a decentralized infrastructure widely used in digital encrypted currencies such as Bitcoin, which can make secure data storage and access possible. Federated learning protects terminal personal data privacy and can carry out efficient machine learning among multiple participants. The numerical results justify that our proposed federated learning and blockchain combination lays a strong foundation for the development of future industrial Internet.
【标题】物联网设备ID和区块链驱动的工业互联网联邦学习
【摘要】随着物联网（IoT）、5G和工业技术的发展，工业互联网已经成为一个新兴的研究领域。由于行业的特殊性，对识别分析服务的时延性、安全性和稳定性提出了更高的要求。传统的域名系统（DNS）由于识别主体形式单一、安全防护意识薄弱，无法满足工业互联网的需求。作为解决方案，本工作将区块链和联邦学习（FL）应用于工业互联网识别。区块链是广泛应用于比特币等数字加密货币的去中心化基础设施，可以使安全的数据存储和访问成为可能。联邦学习保护终端个人数据隐私，可以在多个参与者之间进行高效的机器学习。数值结果证明我们提出的联邦学习和区块链的结合为未来工业互联网的发展奠定了坚实的基础。
评论