An Experiment Study on Federated Learning Testbed

Shen, Cheng Xue, WanliWhile Internet of Things (IoT) can benefit from machine learning by outsourcing model training on the cloud, user data exposure to an untrusted cloud service provider pose threat privacy. Recently, federated is proposed as approach for privacy-preserving (PPML) IoT, while its practicability remains unclear. This work presents evaluation efficiency and privacy performance a readily available framework based PySyft, Python library distributed deep learning. It observed that speed significantly slower centralized due communication overhead. Meanwhile, bears some vulnerability potential man-in-the-middle attacks at network level. The report serves starting point PPML analysis suggests future direction development.