A Quantum Generative Adversarial Network-based Intrusion Detection System

Machine learning has become widely accepted because of its diverse approaches to deal with a variety of cyber security issues. However, their capricious nature of security threats makes classical machine learning cyber systems vulnerable. Moreover, more samples in a big data dataset in classical machine learning approaches could produce the security defence systems weaken. It may create accurate outcomes by processing information which takes longer than expected, or observe poor accuracy because of inefficient training as well as other issues. However, quantum systems have the potential to produce atypical patterns which can not be possible to produce efficiently by classical systems, so we can postulate that quantum computers could use these advantages in that it could outperform the capabilities of classical computers on machine learning tasks. To be specific, an intrusion detection system can detect attack packets or sequence of attack packets at TCP/IP or other protocol level data based on certain patterns present or by profiling to detect anomalies. O(poly(n) gates are required to enable the use of potentially advantageous quantum algorithms with quantum states using he Quantum generative adversarial networks (qGAN) implemented by Qiskit which is a quantum computing tool of IBM.