-
Investigating Application of Deep Neural Networks in Intrusion Detection System Design
Authors:
Mofe O. Jeje
Abstract:
Despite decades of development, existing IDSs still face challenges in improving detection accuracy, evasion, and detection of unknown attacks. To solve these problems, many researchers have focused on designing and developing IDSs that use Deep Neural Networks (DNN) which provides advanced methods of threat investigation and detection. Given this reason, the motivation of this research then, is t…
▽ More
Despite decades of development, existing IDSs still face challenges in improving detection accuracy, evasion, and detection of unknown attacks. To solve these problems, many researchers have focused on designing and developing IDSs that use Deep Neural Networks (DNN) which provides advanced methods of threat investigation and detection. Given this reason, the motivation of this research then, is to learn how effective applications of Deep Neural Networks (DNN) can accurately detect and identify malicious network intrusion, while advancing the frontiers of their optimal potential use in network intrusion detection. Using the ASNM-TUN dataset, the study used a Multilayer Perceptron modeling approach in Deep Neural Network to identify network intrusions, in addition to distinguishing them in terms of legitimate network traffic, direct network attacks, and obfuscated network attacks. To further enhance the speed and efficiency of this DNN solution, a thorough feature selection technique called Forward Feature Selection (FFS), which resulted in a significant reduction in the feature subset, was implemented. Using the Multilayer Perceptron model, test results demonstrate no support for the model to accurately and correctly distinguish the classification of network intrusion.
△ Less
Submitted 26 January, 2025;
originally announced January 2025.
-
Cybersecurity Assessment of Smart Grid Exposure Using a Machine Learning Based Approach
Authors:
Mofe O. Jeje
Abstract:
Given that disturbances to the stable and normal operation of power systems have grown phenomenally, particularly in terms of unauthorized access to confidential and critical data, injection of malicious software, and exploitation of security vulnerabilities in a poorly patched software among others; then developing, as a countermeasure, an assessment solutions with machine learning capabilities t…
▽ More
Given that disturbances to the stable and normal operation of power systems have grown phenomenally, particularly in terms of unauthorized access to confidential and critical data, injection of malicious software, and exploitation of security vulnerabilities in a poorly patched software among others; then developing, as a countermeasure, an assessment solutions with machine learning capabilities to match up in real-time, with the growth and fast pace of these cyber-attacks, is not only critical to the security, reliability and safe operation of power system, but also germane to guaranteeing advanced monitoring and efficient threat detection. Using the Mississippi State University and Oak Ridge National Laboratory dataset, the study used an XGB Classifier modeling approach in machine learning to diagnose and assess power system disturbances, in terms of Attack Events, Natural Events and No-Events. As test results show, the model, in all the three sub-datasets, generally demonstrates good performance on all metrics, as it relates to accurately identifying and classifying all the three power system events.
△ Less
Submitted 23 January, 2025;
originally announced January 2025.
-
UpDown: Programmable fine-grained Events for Scalable Performance on Irregular Applications
Authors:
Andronicus Rajasukumar,
Jiya Su,
Yuqing,
Wang,
Tianshuo Su,
Marziyeh Nourian,
Jose M Monsalve Diaz,
Tianchi Zhang,
Jianru Ding,
Wenyi Wang,
Ziyi Zhang,
Moubarak Jeje,
Henry Hoffmann,
Yanjing Li,
Andrew A. Chien
Abstract:
Applications with irregular data structures, data-dependent control flows and fine-grained data transfers (e.g., real-world graph computations) perform poorly on cache-based systems. We propose the UpDown accelerator that supports fine-grained execution with novel architecture mechanisms - lightweight threading, event-driven scheduling, efficient ultra-short threads, and split-transaction DRAM acc…
▽ More
Applications with irregular data structures, data-dependent control flows and fine-grained data transfers (e.g., real-world graph computations) perform poorly on cache-based systems. We propose the UpDown accelerator that supports fine-grained execution with novel architecture mechanisms - lightweight threading, event-driven scheduling, efficient ultra-short threads, and split-transaction DRAM access with software-controlled synchronization. These hardware primitives support software programmable events, enabling high performance on diverse data structures and algorithms. UpDown also supports scalable performance; hardware replication enables programs to scale up performance. Evaluation results show UpDown's flexibility and scalability enable it to outperform CPUs on graph mining and analytics computations by up to 116-195x geomean speedup and more than 4x speedup over prior accelerators. We show that UpDown generates high memory parallelism (~4.6x over CPU) required for memory intensive graph computations. We present measurements that attribute the performance of UpDown (23x architectural advantage) to its individual architectural mechanisms. Finally, we also analyze the area and power cost of UpDown's mechanisms for software programmability.
△ Less
Submitted 30 July, 2024;
originally announced July 2024.
