AGLCNet: A Novel Attention-Driven GLC Framework for Enhancing IoT Cybersecurity

R. Nandhini, D. Pradeep

2025

Abstract

With the Internet of Things expanding, inter-device security is a sensitive issue owing due to the rising number and sophistication of cyberattacks aimed at the IoT networks. In this paper, novel attention-driven GLC model is proposed to enhance detection of cyberattacks on the IoT environment. The model employs Graph Neural Networks (GNN) to model spatial dependency between devices, Long Short-Term Memory (LSTM) networks for temporal dynamics of the network traffic and Convolutional Neural Networks (CNN) for high level feature extraction. Furthermore, an attention mechanism is added to the architecture to focus on relevant features and device interactions needed to optimally improve the performance of the detection. The model was validated on IoT security benchmark dataset UNSW-NB-15 and outperformed traditional models in terms of accuracy, precision, recall and F1 score. These findings support the model’s ability to provide effective and efficient cyber security for Internet of Things infrastructure. While the model performs well, certain gaps exist in streamlining the computational complexity to enable efficient execution in real-time for low-power IoT devices. Future work will revolve around optimizing the model as well as conducting more experiments on large-scale real-world IoT datasets to evaluate the model’s usability and performance in real-world situations.

Paper Citation

in Harvard Style

Nandhini R. and Pradeep D. (2025). AGLCNet: A Novel Attention-Driven GLC Framework for Enhancing IoT Cybersecurity. In Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - ICRDICCT`25; ISBN 978-989-758-777-1, SciTePress, pages 294-300. DOI: 10.5220/0013927200004919

in Bibtex Style

@conference{icrdicct`2525,
author={R. Nandhini and D. Pradeep},
title={AGLCNet: A Novel Attention-Driven GLC Framework for Enhancing IoT Cybersecurity},
booktitle={Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - ICRDICCT`25},
year={2025},
pages={294-300},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013927200004919},
isbn={978-989-758-777-1},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - ICRDICCT`25
TI - AGLCNet: A Novel Attention-Driven GLC Framework for Enhancing IoT Cybersecurity
SN - 978-989-758-777-1
AU - Nandhini R.
AU - Pradeep D.
PY - 2025
SP - 294
EP - 300
DO - 10.5220/0013927200004919
PB - SciTePress