DIADL: An Energy Efficient Framework for Detecting Intrusion Attack Using Deep LearnIing

In today's era, the Internet's complexity, accessi- bility, and openness has greatly enhanced the security risk of information systems. The Internet's popularity entails many threats of network attacks. Detection of intrusion is one of the major research problems in network security, with the goal of detecting unauthorized access or attacks to secure internal networks. Key ideas are to discover useful patterns or features that characterize a system's user behavior and use the collection of similar features to construct classifiers that can, ideally in real time, detect anomalies and known intrusions. In this proposal, energy efficient framework is designed which detects the intrusion. We predicted the bi- nary and multi class classification by the model with tuned parameters at different levels of learning. The deep learning system enhances the intrusion detection accuracy and provides a new intrusion detection research method. Using a collection of KDD (Knowledge Discovery and Data Mining) benchmark data, we demonstrate that successful and accurate classifiers can be constructed to detect intrusions. In literature, numerous machine learning methods have addressed intrusion detection systems. By analyzing the existing intrusion detection systems, we validated the proposed framework with deep learning. © 2020 ACM.