A Hybrid CNN-LSTM-Attention Framework for Intelligent Cyber Attack Detection with Enhanced Feature Representation

A Hybrid CNN-LSTM-Attention Framework for Intelligent Cyber Attack Detection with Enhanced Feature Representation

mina rajabi1

1) Department of Computer Engineering, Golestan University, Gorgan, Iran, - Email:

IOI Code : AAHD-ABKCAZ
Publication : National Conference on Information Technology, Nanotechnology, Artificial Intelligence and Technological Futures Studies(itnaf.ir)
Abstract :
    The rapid growth of cyber threats has exposed significant limitations in traditional intrusion detection systems, particularly in detecting zero-day attacks and supporting real-time network monitoring. This paper proposes a novel Hybrid CNN-LSTM-Attention (HCLA) framework that integrates Convolutional Neural Networks (CNN), Bidirectional Long Short-Term Memory (BiLSTM), and a Multi-Head Attention mechanism for intelligent cyber attack detection. In the proposed framework, CNN extracts spatial features from network traffic, BiLSTM captures temporal dependencies, and the attention mechanism adaptively emphasizes the most informative features for attack classification. The proposed model was evaluated on the CICIDS2017 dataset. Experimental results demonstrate that HCLA achieves 98.7% accuracy, 98.5% precision, 98.3% recall, and 98.4% F1-score, outperforming conventional CNN, LSTM, GRU, and CNN-LSTM models. Furthermore, the framework achieves an inference latency of 2.3 ms per sample, making it suitable for real-time intrusion detection in high-speed network environments. The results confirm that integrating spatial and temporal feature learning with an attention mechanism significantly improves cyber attack detection performance while maintaining computational efficiency.

Keywords : cyber attack detection deep learning cnn lstm attention mechanism intrusion detection network security