Conditional Generative Adversarial Network for Intrusion Detection System Based on Deep Learning

With the development of big data and artificial intelligence, the tremendous amount of data have led to an increase in the number of cyberattacks. Convolutional neural network combined with intrusion detection system has been widely used to improve the performance of the traditional intrusion detection system. And KDD CPU99 is a frequently used dataset with old traffic data and unbalanced data distribution. In this paper, we proposed an intrusion detection system combined with generative adversarial network and deep residual network. The model is trained with NSL-KDD dataset which is an improved version of KDD CPU99. We also use CTGAN to synthesize data to provide more training samples for unknown cyberattacks. All the data is converted into gray scale images for ResNet-18. As a result, the proposed system perfoms well in experiment and shows great generalizaion performance.