Data Poisoning Attacks on Federated Machine Learning

Abstract

There is a tremendous quantity of data flowing between IoT devices and their users.Some of the material shared in these conversations is very private, including details about specific businesses, military units, or individual patients. As a result, several systems and procedures for security have been proposed. Communication signals these systems' methods of operation may include Elliptic Curve Cryptography, Public Key Infrastructure, and Block chain. Radio frequency identification and a physically unclonable function. To this end, in this article, the benefits and drawbacks of various plans are analyzed. Weaknesses. This data shows that the vast majority of these there are several concerns with the safety, speed, and privacy of protocols.