Open AccessImage Classifiers for Network Intrusions
Author(s) -
David Noever,
Samantha E. Miller Noever
Publication year - 2021
Language(s) - English
Resource type - Conference proceedings
DOI - 10.5121/csit.2021.110504
Subject(s) - computer science , thumbnail , convolutional neural network , artificial intelligence , grayscale , random forest , deep learning , feature (linguistics) , exploit , image (mathematics) , intrusion detection system , machine learning , pattern recognition (psychology) , data mining , computer security , philosophy , linguistics
This research recasts the network attack dataset from UNSW-NB15 as an intrusion detection problem in image space. Using one-hot-encodings, the resulting grayscale thumbnails provide a quarter-million examples for deep learning algorithms. Applying the MobileNetV2’s convolutional neural network architecture, the work demonstrates a 97% accuracy in distinguishing normal and attack traffic. Further class refinements to 9 individual attack families (exploits, worms, shellcodes) show an overall 56% accuracy. Using feature importance rank, a random forest solution on subsets show the most important sourcedestination factors and the least important ones as mainly obscure protocols. The dataset is available on Kaggle.