Published Paper Details:

The increasing sophistication and frequency of cyber-attacks have exposed the limitations of traditional signature-based intrusion detection systems in identifying evolving and previously unseen threats. The presence of highly imbalanced network traffic, scarcity of labeled attack samples, and the dynamic nature of cyber threats significantly reduce the effectiveness of conventional machine learning-based detection approaches. This research proposes an intelligent cyber-attack detection framework that integrates Generative Adversarial Networks (GANs) with advanced Machine Learning (ML) and Deep Learning (DL) techniques to improve the accurate detection and classification of network intrusions. The proposed framework employs GAN-based synthetic data generation to address class imbalance by producing realistic minority attack samples, thereby enhancing the quality and diversity of training datasets. Following data preprocessing and feature engineering, the augmented dataset is utilized to train multiple ML and DL models, including Random Forest, XGBoost, Convolutional Neural Networks (CNN), Long Short-Term Memory (LSTM), and hybrid CNN-LSTM architectures. A comparative evaluation is performed using publicly available benchmark intrusion detection datasets such as CIC-IDS2017, CIC-IDS2018, NSL-KDD, UNSW-NB15, and Bot-IoT. The proposed framework is evaluated using performance metrics including Accuracy, Precision, Recall, F1-Score, Matthews Correlation Coefficient (MCC), Receiver Operating Characteristic-Area under the Curve (ROC-AUC), False Positive Rate (FPR), and computational efficiency. The integration of GAN-based data augmentation with hybrid learning models is expected to substantially improve minority attack detection, reduce false alarms, enhance model generalization, and strengthen robustness against emerging and zero-day cyber threats. The proposed framework provides a scalable, adaptive, and intelligent intrusion detection solution suitable for modern enterprise networks, cloud computing environments, Internet of Things (IoT) ecosystems, and other resource-constrained cyber infrastructures, thereby contributing to the development of next-generation intelligent cybersecurity systems. This work directly supports the objective of developing an integrated GAN-ML-DL framework capable of overcoming class imbalance and limited attack data while achieving highly accurate network intrusion detection.