ML in an Adversarial Environment

Though ML has had many successful real-world applications, including cybersecurity, its methods are vulnerable. Before implementing ML solutions, we must understand how attackers could disrupt them. In this chapter, we will look at some of the vulnerabilities throughout the ML lifecycle. The vulnerabilities are amplified when using ML in an adversarial environment such as cybersecurity. When we implement cybersecurity solutions, whether they leverage ML or not, we know that our adversaries will attempt attacks on them.