An Enhanced Intrusion Detection System using Hybrid CNN-LSTM Architecture

Abstract

The rapid proliferation of the Internet of Things (IoT) and Industrial Internet of Things (IIoT) has revolutionized digital infrastructure but introduced a vast attack surface for sophisticated cyber threats. Traditional Intrusion Detection Systems (IDS), which rely on signature-based matching and manual feature engineering, are increasingly fragile against zero-day exploits and polymorphic malware. This review paper examines the shift toward hybrid deep learning architectures, specifically the integration of Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) networks. By synergistically capturing both spatial intra-flow patterns and long-term temporal dependencies, these hybrid frameworks bridge the semantic gap between raw network telemetry and high-level threat identification, achieving state-of-the-art results across modern benchmark datasets.

Keywords