Enhancing Intrusion Prevention with Explainable Convolutional Neural Networks: Recent Developments and Promising Applications
Ajay Chitnis, Pankaj Ramchandra Chandre, Shafi Pathan · 2023
The proliferation of cyberattacks in recent years has led to an increased demand for effective intrusion prevention systems (IPS). Convolutional Neural Networks (CNNs) have shown promise in improving the accuracy of IPS, but their lack of interpretability hinders their wider adoption. Explainable AI (XAI) techniques have emerged to address this issue, allowing users to understand the reasoning behind a model's predictions. This survey paper provides an overview of recent developments in enhancing IPS with XAI techniques in CNNs. This study reviews state-of-the-art research on integrating interpretability into CNN models, including methods such as attention mechanisms, saliency maps, and LIME. Also, this study discusses about the promising applications of CNN-based IPS with XAI, including network traffic analysis, malware detection, and anomaly detection. Finally, we highlight some of the challenges and opportunities for future research in this area, including the need for standardized evaluation metrics and the potential for combining XAI techniques with other AI methods such as reinforcement learning. Overall, this survey paper aims to provide a comprehensive understanding of the recent advances in using CNNs with XAI for enhancing IPS, and to identify promising directions for future research in this field.