Research Article
Deep Learning-Based Detection of Cyberattacks in Software-Defined Networks
@INPROCEEDINGS{10.1007/978-3-031-36574-4_20, author={Seyed Mohammad Hadi Mirsadeghi and Hayretdin Bahsi and Wissem Inbouli}, title={Deep Learning-Based Detection of Cyberattacks in Software-Defined Networks}, proceedings={Digital Forensics and Cyber Crime. 13th EAI International Conference, ICDF2C 2022, Boston, MA, November 16-18, 2022, Proceedings}, proceedings_a={ICDF2C}, year={2023}, month={7}, keywords={Software-Defined Network Intrusion Detection Deep Learning Dataset Balancing}, doi={10.1007/978-3-031-36574-4_20} }- Seyed Mohammad Hadi Mirsadeghi
Hayretdin Bahsi
Wissem Inbouli
Year: 2023
Deep Learning-Based Detection of Cyberattacks in Software-Defined Networks
ICDF2C
Springer
DOI: 10.1007/978-3-031-36574-4_20
Abstract
This paper presents deep learning models for binary and multiclass intrusion classification problems in Software-defined-networks (SDN). The induced models are evaluated by the state-of-the-art dataset, InSDN. We applied Convolutional Autoencoder (CNN-AE) for high-level feature extraction, and Multi-Layer Perceptron (MLP) for classification that delivers high-performance metrics of F1-score, accuracy and recall compared to similar studies. Highly imbalanced datasets such as InSDN underperform in detecting the instances belonging to the minority class. We use Synthetic Minority Oversampling Technique (SMOTE) to address dataset imbalance and observe a significant detection enhancement in the detection of minority classes.
Copyright © 2022–2025 ICST
