2018/08/16 by Cosimo Ieracitano, Ieracitano, Cosimo, Ahsan Adeel +13
Computer Science · #Advanced Malware Detection Techniques #Anomaly Detection Techniques and Applications #Cryptography and Security (cs.CR) #FOS: Computer and information sciences #I.2.1 #I.5.1 #K.6.5 #Network Security and Intrusion Detection
paper · pdf · doi:10.48550/arxiv.1808.05633
openalex publication_date 2018/08/16 · openalex created_date 2022/08/04 · openalex updated_date 2026/07/28
Attackers have developed ever more sophisticated and intelligent ways to hack\ninformation and communication technology systems. The extent of damage an\nindividual hacker can carry out upon infiltrating a system is well understood.\nA potentially catastrophic scenario can be envisaged where a nation-state\nintercepting encrypted financial data gets hacked. Thus, intelligent\ncybersecurity systems have become inevitably important for improved protection\nagainst malicious threats. However, as malware attacks continue to dramatically\nincrease in volume and complexity, it has become ever more challenging for\ntraditional analytic tools to detect and mitigate threat. Furthermore, a huge\namount of data produced by large networks has made the recognition task even\nmore complicated and challenging. In this work, we propose an innovative\nstatistical analysis driven optimized deep learning system for intrusion\ndetection. The proposed intrusion detection system (IDS) extracts optimized and\nmore correlated features using big data visualization and statistical analysis\nmethods (human-in-the-loop), followed by a deep autoencoder for potential\nthreat detection. Specifically, a pre-processing module eliminates the outliers\nand converts categorical variables into one-hot-encoded vectors. The feature\nextraction module discard features with null values and selects the most\nsignificant features as input to the deep autoencoder model (trained in a\ngreedy-wise manner). The NSL-KDD dataset from the Canadian Institute for\nCybersecurity is used as a benchmark to evaluate the feasibility and\neffectiveness of the proposed architecture. Simulation results demonstrate the\npotential of our proposed system and its outperformance as compared to existing\nstate-of-the-art methods and recently published novel approaches. Ongoing work\nincludes further optimization and real-time evaluation of our proposed IDS.\n