Adopting Edge Computing to Reduce Transmission Latency and Enhance Security in Medical Diagnosis Service

Abstract

Any user may submit the symptoms at any time, anywhere for medical diagnosis as machine learning becomes more prevalent. One reason to use edge computing is to decrease transmission time and latency while providing real-time diagnostic service. However, medical data is exposed when using data-driven machine learning, which must be built over a massive quantity of medical data. Privacy has to be preserved. We propose a lightweight privacy-preserving medical diagnostic method termed LPME, which may help to address the problems described above. The LPME redesigns the Extreme Gradient Building (XGBoost) model based on the edge-cloud model which incorporates encrypted model parameters to reduce quantities of ciphertext calculations to plaintext calculation. Additionally, LPME may offer discreet and fast diagnosis with edge security and privacy protection. Our research shows that LPME's security, efficiency, and efficacy are maximized.