Intrusion Detection in Networks Using Adversarial Networks and Weighted Encoder Components
Nirbhay Kumar Chaubey, Ruby Dahiya, Venkateswaran Radhakrishnan, RVS Praveen, U. Hemavathi, Sangeetha Subramaniam · Advances in information security, privacy, and ethics book series · 2024
This study introduces an innovative approach to network intrusion detection by harnessing the capabilities of adversarial networks and underscoring the importance of encoder components. Conventional intrusion detection systems frequently grapple with the real-time identification of intricate and evolving threats, rendering them susceptible to both false positives and negatives. Our proposed method aims to confront these inherent challenges. The proposed strategy, referred to as ID-ANWE, entails the integration of adversarial networks into the system. This augmentation enhances the system's ability to differentiate between typical network behavior and irregular, potentially malicious activities. Through a training process, these adversarial networks become proficient in distinguishing between legitimate network traffic and potential intrusion attempts, contributing to the establishment of a dynamic and adaptable security framework.