Generative Adversarial Networks for Anomaly Detection in Cyber Security: A Review

The CIA model is the basis for cybersecurity in its broadest sense; this model prioritizes maintaining the confidentiality, integrity, and availability of data. The majority of attacks are launched by the attackers by exploiting flaws in the communication protocols. Because these assaults put the reputations of service providers in jeopardy, more effective mitigation strategies are essential for mitigating attacks. Therefore, there is major concern for service providers. Generative adversarial networks (GANs) are emerging techniques that are gaining importance in AI-based security defense systems. GANs are being used by the security industry to great effect in areas like intrusion monitoring or detection, steganography, cryptography, password cracking, and anomaly detection, to name a few. This study presents a comprehensive literature overview of the use of GANs in cybersecurity, including an in-depth examination of the most popular stable cybersecurity datasets in use today and the specific extended GAN frameworks behind their creation.