Secure federated learning architecture for fuzzy classifier in healthcare environment

【Author】 Vishwakarma, Santosh; Goswami, Rajat Subhra; Nayudu, P. Prathap; Sekhar, Krovi Raja; Arnepalli, Pandu Ranga Rao; Thatikonda, Ramya; Abdel-Rehim, Wael M. F.

【Source】SOFT COMPUTING

【影响因子】3.732

【Abstract】There have been many developments in recent years based on the Internet of Things (IoT), particularly for managing data pertaining to healthcare as it is now known as the Internet of Health Things (IoHT). As those healthcare data ensure improved accuracy, security with enhanced integrity, and better quality data, some these data must be provisioned to such smart healthcare, which is a time-consuming problem in this organization. To improve data provisioning, data must be more secure and privacy protected, as enabled by federated learning and privacy policies. If the data is private, it is possible to learn about it with the owner's consent. With recent advancements, data process flow related to IoHT is associated with various IoT devices as edge nodes. As there is a problem with a partial level of trained nodes, learning at the nodes is much more difficult as it requires a fully decentralized environment, improved and trained datasets, data provisioning, and security. In this paper, a Hy-FL-based Blockchain approach is suggested because it can manage trust and trained data based on federated learning with better authentication thanks to blockchain technology. This proposed approach enables the encryption of trained data on federated nodes and aggregated data. In the analysis, IoHT-based data manageability is handled safely in terms of energy use, data accuracy, predicated value, etc.

【Keywords】Federated learning; Data privacy; Security; Authorization; Internet of medical things

【发表时间】2023 2023 JUL 11

【收录时间】2023-08-06

【文献类型】实验仿真

【主题类别】

区块链应用-实体经济-医疗领域