Downloads: 5 | Views: 92 | Weekly Hits: ⮙1 | Monthly Hits: ⮙1
Analysis Study Research Paper | Computer Engineering | India | Volume 10 Issue 3, March 2021 | Rating: 4.7 / 10
Privacy-Preserving AI at the Edge: Techniques and Applications
Venkata Naga Sai Kiran Challa [7]
Abstract: IOT sensors, smartphones, cars, etc., are the popularly used EC devices that facilitate local processing of data, which has given rise to privacy concerns. The following is a list of privacy - preserving AI techniques at the edge that tackles these problems to ensure data privacy as much as AI performance: Subsequently, it discusses the crucial privacy - preserving technologies for edge AI, specifically FL, HE, SMPC, and DP. Thus it discusses their operations, their uses and the advantages they have over the traditional centralised models. This paper also explains the present scenarios related to the use of these techniques in different domains like health care, smart home, self - driving cars, and industrial IoT domains. Specifically, insights for the development of privacy - preserving AI at the edge of the next years are revealed as well as 5G, advanced hardware, and regulation perspectives.
Keywords: Artificial intelligence at the edge, distributed learning, cryptography, secure multi - party computation, differential privacy, privacy - enhancing technologies, real - time data analysis, decentralized processing, health care, internet of things, self - driving cars
Edition: Volume 10 Issue 3, March 2021,
Pages: 2013 - 2023