Hybrid Deep Learning–Based Network Intrusion Detection Using ConvLSTM, Spiking Neural Networks and Digital Twin Architecture

Read full text Summarise with VIRAT
Authors: Y. Chiranjeevi Ashok Kumar, D. Srija
Cited by: 0
References: 3

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Network Intrusion Detection is a pressing issue for today’s enterprise systems as the number of network attacks is ever increasing. Traditional machine learning (ML) solutions to this problem are hampered by their inability to prevent false alarms and analyse the temporal behaviour of system activities. While deep learning (DL) solutions are shown to achieve better accuracy than their traditional ML counterparts, they tend to lack the required optimisation in terms of balancing spatiotemporal feature extraction, real-time processing and energy efficiency. Traditional machine learning-based Network Intrusion Detection System (NIDS) face lots of challenges, such as long training time, large amounts of data, and high dimensionality of network data. Meanwhile, Deep learning (DL) has become a trending topic in recent years due to its better performance and more outstanding features in data analysis. Therefore, a hybrid deep learning based Network Intrusion Detection System (NIDS) model that improves the performance of the Network Intrusion Detection System by using the strengths of various models was proposed in this paper. In this proposed model, the data in the network packet is fed into the ConvLSTM module to learn the features of the spatiotemporal patterns of the network traffic. The extracted feature data is then input into the Spiking Neural Networks (SNN) to detect the threats at an energy-efficient temporal scale. The Digital Twin (DT) of the NIDS system is also incorporated with the proposed model for contextual threat scoring and further improves the effectiveness of the proposed hybrid deep learning-based model. Experimental results using the CIC-IDS2017 and UNSW-NB15 datasets showed higher accuracy, fewer false alarms, and lower detection time compared to other systems, demonstrating the practicability of the proposed hybrid deep learning-based NIDS model in the massive real-time cybersecurity field.

Fields of Study & Subjects

Network Security and Intrusion Detection Network Packet Processing and Optimization Internet Traffic Analysis and Secure E-voting

Publication details

Journal / source
Research Digest on Engineering Management and Social Innovations
Publication date
31 March 2026
Language
EN