边缘学习:隐私计算架构、技术现状与展望

Edge Learning:Privacy Computing Architecture Key Technologies and Challenges

边缘学习旨在实现云-边-端协同的机器学习模型训练和预测,天然具有一定隐私保护能力。但是,边缘学习过程面临新的安全与隐私泄露风险。为此,本文从边缘学习的概念出发,重点围绕边缘学习安全与隐私泄露风险及其隐私计算架构、关键技术、未来方向展开论述。

Edge learning is mainly applicable in collaborativemachine learning and model prediction scenarios that involve cloud-edge-end architecture.This distributed nature of edge learning naturally provides a certain level of privacy protection.However,collaborative learning faces some new privacy risks that must be addressed.Therefore,this paper explores the concept of edge learning and focuses on the security and privacy disclosure risks associated with it.Additionally,the paper delves into the technical architecture,key technologies,and future directions of privacy computing in edge learning.