The research presents a three-stage training model for intrusion detection in edge computing devices. The model uses advanced deep learning techniques and is designed to work effectively within resource-constrained devices. The training model is enhanced with a pruning methodology and model compression techniques to maintain high accuracy levels. The model was tested on the UNSW-NB15 dataset and showed a significant reduction in model dimensions while maintaining accuracy. The study aims to balance intrusion detection effectiveness and resource consumption.
Publication date: 1 Feb 2024
Project Page: https://doi.org/10.5121/ijcnc.2024.16102
Paper: https://arxiv.org/pdf/2401.17546
